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 2161 - (view) (download)

1 : dbm 570 This is the HISTORY file for the Yale SML/NJ CVS repository.
2 :    
3 :     An entry should be made for _every_ commit to the repository.
4 :     The entries in this file will be used when creating the README
5 :     for new versions, so keep that in mind when writing the
6 :     description.
7 :    
8 :     The form of an entry should be:
9 :    
10 : mblume 1931
11 :     Name: Your Name (web page or e-mail)
12 : blume 2155 Date: yyyy/mm/dd hh:mm:ss
13 : mblume 1931 Tag: cvs-tag-for-this-change
14 : mblume 1480 Description:
15 :    
16 :     ----------------------------------------------------------------------
17 : mblume 1923 Name: Matthias Blume (http://www.tti-c.org/blume/)
18 : blume 2155 Date: 2006/11/02 12:20:00 CST
19 :     Description:
20 :    
21 :     Reorganized directory layout.
22 :    
23 :     This is a temporary solution, more reorganization is to come.
24 :    
25 :     The basic idea is to have a number of toplevel trees, each corresponding
26 :     to a well-defined part of the overall system. Each part can be
27 :     maintained individually, even in separate source repositories,
28 :     although currently we still serve everything out of the main
29 :     smlnj-gforge tree.
30 :    
31 :     Installer and and scripts have been updated to reflect the new
32 :     layout. The installer (base/src/system/smlnj/installer) is now
33 :     "scriptable" to avoid burning too much knowledge about the
34 :     layout into SML source code. The main script used by the installer
35 :     is in config/actions.
36 :    
37 :     The main change is that many of the subdirectories of what used to
38 :     be known as "src" have moved to the toplevel. The "src" directory
39 :     itself has moved down into a subtree called "base". (We may eventually
40 :     get rid of the extra level of indirection represented by "src".)
41 :    
42 :     The layout is now as follows:
43 :    
44 :     toplevel tree name | default repository
45 :     | (using gf=svn://smlnj-gforge.cs.uchicago.edu/smlnj)
46 :     -------------------+----------------------------------------------------
47 :     config | $gf/config/trunk
48 :     |
49 :     base | $gf/sml/trunk
50 :     |
51 :     smlnj-lib | $gf/smlnj-lib/trunk
52 :     MLRISC | $gf/MLRISC/trunk
53 :     |
54 :     ml-yacc | $gf/ml-yacc/trunk
55 :     ml-lex | $gf/ml-lex/trunk
56 :     ml-burg | $gf/ml-burg/trunk
57 :     lexgen | $gf/lexgen/trunk
58 :     heap2asm | $gf/heap2asm/trunk
59 :     |
60 :     cml | $gf/cml/trunk
61 :     eXene | $gf/eXene/trunk
62 :     ckit | $gf/ckit/trunk
63 :     nlffi | $gf/nlffi/trunk
64 :     |
65 :     smlnj-c | $gf/smlnj-c/trunk
66 :    
67 :     In $gf/admin there are a few useful shell scripts for checking out
68 :     and maintaining the entire collection of trees:
69 :    
70 :     admin/checkout-all.sh [dir]
71 :     -- optionally creates dir and checks out all trees from their
72 :     default repositories; if dir is missing, checkout into .
73 :     admin/refresh-all.sh [dir]
74 :     -- looks at all trees (from the above list) in dir (default: .)
75 :     and runs "svn update" if the tree exists and is under
76 :     subversion control;
77 :     non-existing or non-subversion trees are skipped
78 :     admin/stat-all.sh [dir]
79 :     -- like refresh-all.sh, but runs "svn stat" instead
80 :    
81 :     ----------------------------------------------------------------------
82 :     Name: Matthias Blume (http://www.tti-c.org/blume/)
83 :     Date: 2006/10/05 11:30:00 CDT
84 : mblume 2066 Tag: blume-20061005-merge-amd64
85 :     Description:
86 :    
87 :     Merged code for AMD64 backend (Mike Rainey's work).
88 :     Everything is hooked up but untested.
89 :    
90 :     ----------------------------------------------------------------------
91 :     Name: Matthias Blume (http://www.tti-c.org/blume/)
92 : mblume 1931 Date: 2006/05/17 14:00:00 CST
93 :     Tag: Release_110_59
94 :     Description:
95 :    
96 :     New working version (110.58). NEW BOOTFILES!
97 :    
98 :     I am freezing 110.59. Changes other than the version-number increase:
99 :    
100 :     eXene:
101 :    
102 :     * committed changes to eXene from Alley Stoughton:
103 :     "fixed bugs in X authorization and resource handling, as well
104 :     as in the pile and viewport widgets"
105 :    
106 :     Runtime:
107 :     * fixed linking problem with NetBSD 3.x.
108 :    
109 :     Lexgen:
110 :     * lexgen tool handles non-ascii characters in 7-bit mode the same
111 :     way that ml-lex does
112 :    
113 :     * lexgen propagates exceptions the same way that ml-lex does
114 :    
115 :     CML:
116 :     * Fixed a bug in the SyncVar polling functions (iGetPoll,
117 :     mTakePoll, and mGetPoll) that could lead to livelock.
118 :    
119 :     ----------------------------------------------------------------------
120 :     Name: Matthias Blume (http://www.tti-c.org/blume/)
121 : mblume 1926 Date: 2006/05/12 15:20:00 CST
122 :     Tag: blume-20060515-makedepend
123 :     Description:
124 :    
125 :     Implemented ml-makedepend (i.e., CM.sources) in a better (more robust)
126 :     way. This should hopefully fix the ml-makedepend problem permanently.
127 :    
128 :     ----------------------------------------------------------------------
129 :     Name: Matthias Blume (http://www.tti-c.org/blume/)
130 : mblume 1925 Date: 2006/05/12 15:15:00 CST
131 :     Tag: blume-20060512-makedepend
132 :     Description:
133 :    
134 :     Fixed long-standing bug with ml-makedepend where it would output a
135 :     spurious dependency to a non-existing file. (This is a simple fix.
136 :     It might need further looking into.)
137 :    
138 :     ----------------------------------------------------------------------
139 :     Name: Matthias Blume (http://www.tti-c.org/blume/)
140 : mblume 1923 Date: 2006/04/20 10:30:00 CDT
141 :     Tag: blume-20060420-heap2exec
142 :     Description:
143 :    
144 :     Committed patches received from Johannes 5 Joemann (joemann@befree.free.de)
145 :     that enable heap2exec under Linux and FreeBSD.
146 :    
147 :     ----------------------------------------------------------------------
148 : jhr 1921 Name: John Reppy
149 :     Date: 2006/04/14
150 : blume 2155 Tag:
151 : jhr 1921 Description:
152 :    
153 :     MLRISC changes:
154 :     1) renamed GAS_PSEUDO_OPS to AS_PSEUDO_OPS and put it in its own file.
155 :     2) added support for NOTB and XORB operators in pseudo-op expressions
156 :     3) added DarwinPseudoOp functor that supports Darwin's assembler syntax.
157 :     4) added support for 64-bit integer literals
158 :    
159 :     ----------------------------------------------------------------------
160 : jhr 1922 Name: Matthias Blume (http://www.tti-c.org/blume/)
161 : mblume 1907 Date: yyyy/mm/dd
162 :     Tag: Release_110_58
163 :     Description:
164 :    
165 :     New working version (110.58). NEW BOOTFILES!
166 :    
167 :     ----------------------------------------------------------------------
168 : mblume 1896 Name: Matthias Blume (http://www.tti-c.org/blume/)
169 : mblume 1905 Date: 2006/03/01 22:15:00 CST
170 :     Tag: blume-20060301-lexgen
171 :     Description:
172 :    
173 :     Incorporated several bugfixes to lexgen. Compiler now compiles to
174 :     fixpoint when using lexgen instead of ml-lex.
175 :    
176 :     ----------------------------------------------------------------------
177 :     Name: Matthias Blume (http://www.tti-c.org/blume/)
178 : mblume 1902 Date: 2006/02/28 22:45:00 CST
179 :     Tag: blume-20060228-lexgen
180 :     Description:
181 :    
182 :     Removed ml-flex and added lexgen instead, using Aaron Turon's newly
183 :     provided tarball. The generated lexers still have problems.
184 :    
185 :     ----------------------------------------------------------------------
186 :     Name: Matthias Blume (http://www.tti-c.org/blume/)
187 : mblume 1898 Date: 2006/02/26 23:30:00 CST
188 :     Tag: blume-20060226-ml-flex
189 :     Description:
190 :    
191 :     Removed ml-flex' dependency on regexp-lib.cm. Turned installation
192 :     of ml-flex on by default.
193 :    
194 :     ----------------------------------------------------------------------
195 :     Name: Matthias Blume (http://www.tti-c.org/blume/)
196 : mblume 1896 Date: 2006/02/24 16:30:00 CST
197 :     Tag: blume-20060224-ml-flex
198 :     Description:
199 :    
200 :     Added ml-flex sources. Partially integrated, but should not be
201 :     turned on yet! (Read: leave it commented-out in config/targets!)
202 :    
203 :     ----------------------------------------------------------------------
204 : jhr 1893 Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
205 :     Date: 2006/02/23
206 :     Tag: jhr-20060223-x86-64
207 :     Description:
208 :     Changes to support building on x86-64 systems (using the 32-bit mode).
209 :     Also cleaned up signal handling on Linux. Support for pre-2.2 Linux
210 :     kernels dropped.
211 :    
212 :     ----------------------------------------------------------------------
213 : allenleung 1890 Name: Allen Leung (leunga (at) reservoir (dot) com)
214 :     Date: 2006/02/22 15:07:00 EST
215 :     Tag: leunga-20060222-mdl-gen
216 :     Description:
217 :    
218 :     CM has changed. Updated the script for rebuilding the MLRISC generated
219 :     files.
220 :    
221 :     ----------------------------------------------------------------------
222 : mblume 1886 Name: Matthias Blume (http://www.tti-c.org/blume/)
223 :     Date: 2006/02/14 17:30:00 CST
224 :     Tag: blume-20060214-intel-mac
225 :     Description:
226 :    
227 :     Hooked code for Darwin-specific Intel ABI into main compiler.
228 :     (This is a temporary solution which relies on the fact that the
229 :     compiler itself does not use NLFFI. Eventually we need to divorce
230 :     intel mac from generic x86 unix code and make separate sets of binaries.)
231 :    
232 :     ----------------------------------------------------------------------
233 : jhr 1882 Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
234 : jhr 1885 Date: 2006/02/13
235 :     Tag: jhr-20060213-intel-mac
236 :     Description:
237 :     Changed MLRISC x86 CCalls for partial support of MacOS X ABI.
238 :    
239 :     ----------------------------------------------------------------------
240 :     Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
241 : jhr 1882 Date: 2006/02/06
242 :     Tag: jhr-20060206-intel-mac
243 :     Description:
244 :     Changes to support MacOS X on Intel hardware. The C-calls support in
245 :     MLRISC must also be updated to support the MacOS X ABI.
246 :    
247 :     ----------------------------------------------------------------------
248 : mblume 1863 Name: Matthias Blume (http://www.tti-c.org/blume/)
249 : mblume 1879 Date: 2005/12/16 00:15:00 CST
250 :     Tag: blume-20051216-cm
251 :     Description:
252 :    
253 :     Improved error reporting and handling in CM.
254 :    
255 :     ----------------------------------------------------------------------
256 :     Name: Matthias Blume (http://www.tti-c.org/blume/)
257 : mblume 1878 Date: 2005/11/21 14:00:00 CST
258 :     Tag: blume-20051121-html-lib
259 :     Description:
260 :    
261 :     Enabling $/html-lib.cm in config/preloads did not work. This is fixed
262 :     now. (Since the anchor mapping for html-lib.cm is not yet in effect
263 : blume 2155 at the time when config/preloads is processed, the library has to
264 : mblume 1878 be referred to by another name. In 110.57 this would be
265 :     $SMLNJ-LIB/HTML/html-lib.cm. I arranged for $smlnj/smlnj-lib/html-lib.cm
266 :     to be valid as well.)
267 :    
268 :     Thanks to Todd Wilson (Fresno) for alerting me to this issue.
269 :    
270 :     ----------------------------------------------------------------------
271 :     Name: Matthias Blume (http://www.tti-c.org/blume/)
272 : mblume 1874 Date: 2005/11/19 13:45:00 CST
273 :     Tag: Release_110_57
274 :     Description:
275 :    
276 :     Moved release tag for 110.57.
277 :    
278 :     ----------------------------------------------------------------------
279 :     Name: Matthias Blume (http://www.tti-c.org/blume/)
280 : mblume 1873 Date: 2005/11/19 23:35:00 CST
281 :     Tag: blume-20051119-fish
282 :     Description:
283 :    
284 :     Fixed a problem in config/install.sh where it tries to "fish" the
285 :     name of the CM metadata directory from the wrong place (because
286 :     the physical location of basis.cm has changed). Also, corrected the
287 :     path anchor for $/html-lib.cm. (Thanks to M. Fluet for pointing out
288 :     these problems.)
289 :    
290 :     ----------------------------------------------------------------------
291 :     Name: Matthias Blume (http://www.tti-c.org/blume/)
292 : mblume 1867 Date: 2005/11/16 14:30:00 CST
293 : mblume 1874 Tag: blume-orig-110_57
294 : mblume 1867 Description:
295 :    
296 :     New working version (110.57). NEW BOOTFILES!
297 :    
298 : mblume 1874 UPDATE: The release tag has been moved to include a fix for a brown-paper-
299 :     bag glitch (see above).
300 :    
301 : mblume 1867 ----------------------------------------------------------------------
302 :     Name: Matthias Blume (http://www.tti-c.org/blume/)
303 : mblume 1866 Date: 2005/11/16 10:25:00 CST
304 :     Tag: blume-22051116-btrace
305 :     Description:
306 :    
307 :     Fixed problem with bogus exception message when using back-trace
308 :     facility.
309 :    
310 :     ----------------------------------------------------------------------
311 :     Name: Matthias Blume (http://www.tti-c.org/blume/)
312 : mblume 1865 Date: 2005/11/15 22:35:00 CST
313 :     Tag: blume-22051115-array2copy
314 :     Description:
315 :    
316 :     Added simple implementation of Array2.copy. (Warning: mostly untested.)
317 :    
318 :     ----------------------------------------------------------------------
319 :     Name: Matthias Blume (http://www.tti-c.org/blume/)
320 : mblume 1864 Date: 2005/11/15 21:35:00 CST
321 :     Tag: blume-22051115-versiontool
322 :     Description:
323 :    
324 :     Reversed change to src/system/smlnj/internals/versiontool.cm. This
325 :     file gets loaded as a tool -- by the equivalent of CM.make during the
326 :     run of CMB.make. Thus, CMB's path configuration is meaningless for
327 :     it. Instead, it has the status of "user code", so it should use
328 :     $/basis.cm to refer to the Basis library. (At least that's true for
329 :     the purpose of bootstrapping the previous change. In the future it
330 :     might make sense to have versiontool.cm refer to
331 :     $smlnj/basis/basis.cm, i.e., the version of the Basis that the
332 :     compiler itself uses.)
333 :    
334 :     Also patched src/system/testml to have it activate those extra anchor
335 :     bindings in config/extrapathconfig.
336 :    
337 :     ----------------------------------------------------------------------
338 :     Name: Matthias Blume (http://www.tti-c.org/blume/)
339 : mblume 1863 Date: 2005/11/15 16:35:00 CST
340 :     Tag: blume-22051115-libraries
341 :     Description:
342 :    
343 :     This change affects the way the following libraries are tied into
344 :     the system:
345 :    
346 :     $/basis.cm
347 :     $/smlnj-lib.cm
348 :     $/pp-lib.cm
349 :     $/controls-lib.cm
350 :     $/html-lib.cm
351 :     $/ml-yacc-lib.cm
352 :    
353 :     These libraries are now internally (as seen from the source code
354 :     of the implementation itself) known by the following names:
355 :    
356 :     $smlnj/basis/basis.cm
357 :     $smlnj/smlnj-lib/smlnj-lib.cm
358 :     $smlnj/smlnj-lib/pp-lib.cm
359 :     $smlnj/smlnj-lib/controls-lib.cm
360 :     $smlnj/smlnj-lib/html-lib.cm
361 :     $smlnj/ml-yacc/ml-yacc-lib.cm
362 :    
363 :     This makes it possible to work with code that requires different
364 :     versions of these libraries, and which refers to these libraries using
365 :     their "default" names (i.e., the first set of names above). In other
366 :     words, one can un-define or re-define those default names without
367 :     compromising the proper functioning of the compiler itself.
368 :    
369 :     A similar procedure had already been performed for several of the
370 :     MLRISC libraries that are linked into the compiler. I did some
371 :     cleanup on this code.
372 :    
373 :     A new file in the config directory (named extrapathconfig) is
374 :     responsible for setting up path anchors that the compiler itself does
375 :     not need, but that are typically required by user code.
376 :    
377 :     ----------------------------------------------------------------------
378 : mblume 1862 Name: Matthias Blume
379 :     Date: 2005/11/07 17:05:00 CST
380 :     Tag: blume-20051107-slice
381 :     Description:
382 :    
383 :     Fixed erroneous out-of-bounds test in the "update" function of
384 :     various *ArraySlice modules. (Thanks to Vesa A. Norrman for pointing
385 :     out the problem.)
386 :    
387 :     Pushed some Basis changes through ML-Lex, CML, and eXene.
388 :    
389 :     ----------------------------------------------------------------------
390 : jhr 1858 Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
391 : jhr 1861 Date: 2005/11/07
392 :     Description:
393 :     Fixed a Basis incompatibility: The depreciated function Substring.all
394 :     was removed (use Substring.full instead).
395 :    
396 :     ----------------------------------------------------------------------
397 :     Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
398 : jhr 1858 Date: 2005/11/05
399 :     Description:
400 :     Tweaked interval set API in SML/NJ library; see the CHANGES file for details.
401 :    
402 :     ----------------------------------------------------------------------
403 : mblume 1856 Name: Matthias Blume (http://www.tti-c.org/blume/)
404 :     Date: 2005/11/02 16:25:00 CST
405 :     Tag: blume-20051102-hex
406 :     Description:
407 :    
408 :     * Runtime system bootstrap code now accepts hex digits in BOOTLIST
409 :     in either upper case or lower case format.
410 :     * Pushed changes to names of Pack<N>{Big,Little} structures through
411 :     CML and eXene.
412 :    
413 :     ----------------------------------------------------------------------
414 : jhr 1852 Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
415 : jhr 1855 Date: 2005/11/02
416 :     Description:
417 :     Fixed a Basis incompatibility: Pack<N>{Big,Little} structures should
418 :     be called PackWord<N>{Big,Little}.
419 :    
420 :     ----------------------------------------------------------------------
421 :     Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
422 : jhr 1852 Date: 2005/10/28
423 :     Description:
424 :     Fixed a minor Basis incompatibility: hex digits should be upper case.
425 :    
426 :     ----------------------------------------------------------------------
427 : mblume 1846 Name: Matthias Blume (blume (at) tti - c (dot) org)
428 :     Date: 2005/10/25 16:00:00 CDT
429 :     Tag: Release_110_56
430 :     Description:
431 :    
432 :     New working version (110.56). NEW BOOTFILES!
433 :    
434 :     ----------------------------------------------------------------------
435 : jhr 1838 Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
436 : jhr 1840 Date: 2005/10/25
437 :     Description:
438 :     Added interval sets to utility library (signatures INTERVAL_DOMAIN
439 :     and INTERVAL_SET, and functor IntervalSetFn).
440 :    
441 :     ----------------------------------------------------------------------
442 :     Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
443 : jhr 1838 Date: 2005/10/14
444 :     Description:
445 :     Add Zhong Shao's fix for datatype equality functions.
446 :    
447 :     ----------------------------------------------------------------------
448 : allenleung 1836 Name: Allen Leung (leunga (at) reservoir (dot) com)
449 : allenleung 1837 Date: 2005/10/14 12:30:00 EST
450 :     Tag: leunga-20051014-x86-spill
451 :     Description:
452 :    
453 :     Bug fix a bug found by Carl Hauser.
454 :     There was a typo in the reload code for FCMP in x86SpillInstr.sml.
455 :    
456 :     ----------------------------------------------------------------------
457 :     Name: Allen Leung (leunga (at) reservoir (dot) com)
458 : allenleung 1836 Date: 2005/10/14 11:15:00 EST
459 :     Tag: leunga-20051014-x86-asm
460 :     Description:
461 :    
462 :     Removed some debugging code in file x86Asm.sml.
463 :     The function emit_operand was printing out debugging output.
464 :    
465 :     ----------------------------------------------------------------------
466 : jhr 1813 Name: John Reppy
467 : jhr 1831 Date: 2005/07/27
468 :     Tag:
469 :     Description:
470 :    
471 : jhr 1832 Fixed ml-lex to recognize "\r" as representing carriage return.
472 :    
473 :     ----------------------------------------------------------------------
474 :     Name: John Reppy
475 :     Date: 2005/07/27
476 :     Tag:
477 :     Description:
478 :    
479 : jhr 1831 Fixed ml-yacc to work on files with non-native end-of-line
480 :     encodings (e.g., Windows text file on a Unix system).
481 :    
482 :     ----------------------------------------------------------------------
483 :     Name: John Reppy
484 : jhr 1824 Date: 2005/07/20
485 :     Tag:
486 :     Description:
487 :    
488 :     Added changes from Dominic Evans (oldmanuk (at) gmail (dot) com)
489 :     to support HPUX 11.
490 :    
491 :     ----------------------------------------------------------------------
492 :     Name: John Reppy
493 : jhr 1813 Date: 2005/07/06
494 :     Tag:
495 :     Description:
496 :    
497 :     Changes to the SML/NJ library. See smlnj-lib/CHANGES for details.
498 :    
499 :     ----------------------------------------------------------------------
500 : mblume 1808 Name: Matthias Blume (blume (at) tti - c (dot) org)
501 :     Date: 2005/07/06 00:45:00 CDT
502 :     Tag: blume-20050706-slice-copy
503 :     Description:
504 :    
505 :     Fixed reversed logic for deciding whether to "copy up" or "copy down"
506 :     in *-array-slice.sml.
507 :    
508 :     ----------------------------------------------------------------------
509 : allenleung 1806 Name: Allen Leung (leunga (at) reservoir (dot) com)
510 : allenleung 1807 Date: 2005/05/31 17:00:00 EST
511 :     Tag: leunga-20050531-cygwin-fault-2
512 :     Description:
513 :    
514 :     A typo in the cygwin code fixed.
515 :    
516 :     ----------------------------------------------------------------------
517 :     Name: Allen Leung (leunga (at) reservoir (dot) com)
518 : allenleung 1806 Date: 2005/05/31 16:47:00 EST
519 :     Tag: leunga-20050531-cygwin-fault
520 :     Description:
521 :    
522 : blume 2155 Updated Cygwin's fault/signal handling to match the Windows version.
523 : allenleung 1806 Updated the export list.
524 :    
525 :     ----------------------------------------------------------------------
526 : mblume 1791 Name: Matthias Blume (blume (at) tti - c (dot) org)
527 : mblume 1795 Date: 2005/05/18 13:35:00 CDT
528 :     Tag: Release_110_54
529 :     Description:
530 :    
531 :     New working version (110.54). NEW BOOTFILES!
532 :    
533 :     ----------------------------------------------------------------------
534 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
535 : mblume 1794 Date: 2005/05/18 11:58:00 CDT
536 :     Tag: blume-20050518-installer
537 :     Description:
538 :    
539 :     Added support scripts for Mac OS X PackageMaker and modified
540 :     config/install.sh so that it supports re-dumping a heap image
541 :     after customization.
542 :    
543 :     ----------------------------------------------------------------------
544 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
545 : mblume 1793 Date: 2005/05/18 10:55:00 CDT
546 :     Tag: blume-20050518-realdiv-noovld
547 :     Description:
548 :    
549 :     Un-overloaded / to work around bug in overloading resolution code.
550 :    
551 :     ----------------------------------------------------------------------
552 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
553 : mblume 1792 Date: 2005/05/16 23:50:00 CDT
554 :     Tag: blume-20050516-redump-heap
555 :     Description:
556 :    
557 :     Added mechanism for re-creating a heap file for the interactive system
558 :     after configuration variables have been changed.
559 :    
560 :     CM.redump_heap : string -> unit
561 :    
562 :     This is much like SMLofNJ.exportML, but starting from the resulting
563 : blume 2155 heap does not return to the caller of CM.redump_heap but
564 : mblume 1792 restarts the interactive system from scratch. The original call of
565 :     CM.redump_heap does not return but ends the interactive session.
566 :     Thus, CM.redump_heap is a lot like SMLofNJ.exportFn.
567 :    
568 :     Internally, redump_heap winds the dynamic execution context back to
569 :     the point where the original heap image was created and re-executes
570 :     the heap image generation code in the boot code.
571 :    
572 :     ----------------------------------------------------------------------
573 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
574 : mblume 1791 Date: 2005/05/09 21:55:00 CDT
575 :     Tag: blume-20050509-word64
576 :     Description:
577 :    
578 :     Added a hack to the existing hack known as Word64 to make fromString
579 :     behave correctly. I am still not sure whether Word64.scan will work
580 :     as specified with respect to the interaction of radix and prefix.
581 :    
582 :     ----------------------------------------------------------------------
583 : allenleung 1785 Name: Allen Leung (leunga (at) reservoir (dot) com)
584 :     Date: 2005/05/04 11:50:00 EST
585 :     Tag: leunga-20050504-checkgc
586 :     Description:
587 :    
588 : blume 2155 Added a gc protocol checking phase. This phase is enabled with
589 : allenleung 1785 the flag "check-gc". "debug-check-gc" turns on the verbose mode.
590 :    
591 :     ----------------------------------------------------------------------
592 : mblume 1741 Name: Matthias Blume (blume (at) tti - c (dot) org)
593 : mblume 1784 Date: 2005/05/04 10:45:00 CDT
594 :     Tag: blume-20050504-intinf
595 :     Description:
596 :    
597 :     Fixed a bug in the implementation of div and mod for IntInf.
598 :     Thanks to Neophytos Michael for reporting the problem.
599 :    
600 :     ----------------------------------------------------------------------
601 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
602 : mblume 1783 Date: 2005/05/04 10:35:00 CDT
603 :     Tag: blume-20050504-join
604 :     Description:
605 :    
606 :     Added a "join" combinator to the ParserComb module in smlnj-lib.cm.
607 :    
608 :     ----------------------------------------------------------------------
609 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
610 : mblume 1770 Date: 2005/02/28 23:40:00 CST
611 :     Tag: blume-20050228-mVar
612 :     Description:
613 :    
614 :     Fixed serious bug (brown paper bag variety) in new implementation of
615 :     structure Atom in CML. (I had accidentally used a mailbox instead of
616 :     an mvar, leaving the door open for races.)
617 :    
618 :     ----------------------------------------------------------------------
619 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
620 : mblume 1761 Date: 2005/02/25 15:00:00 CST
621 :     Tag: Release_110_53
622 :     Description:
623 :    
624 :     New working version (110.53). NEW BOOTFILES!
625 :    
626 :     ----------------------------------------------------------------------
627 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
628 : mblume 1760 Date: 2005/02/25 14:50:00 CST
629 :     Tag: blume-20050225-susp
630 :     Description:
631 :    
632 :     Brought back SMLofNJ.Susp. The underlying suspension type is the one
633 :     implemented in Core -- which means that it is the same as the one used
634 :     by the lazy extension.
635 :    
636 :     ----------------------------------------------------------------------
637 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
638 : mblume 1759 Date: 2005/02/24 16:50:00 CST
639 :     Tag: blume-20050224-cml-atom
640 :     Description:
641 :    
642 :     Simpler and at the same time more general implementation of structure
643 :     Atom in CML.
644 :    
645 :     ----------------------------------------------------------------------
646 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
647 : mblume 1758 Date: 2005/02/15 17:35:00 CST
648 :     Tag: blume-20050215-tools
649 :     Description:
650 :    
651 :     Created new "tools" directory under "src" and moved "TraceDebugProf"
652 :     there.
653 :    
654 :     ----------------------------------------------------------------------
655 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
656 : mblume 1755 Date: 2005/02/10 17:55:00 CST
657 :     Tag: blume-20050210-longlong
658 :     Description:
659 :    
660 :     Implemented "long long" arguments and results for NLFFI. (Only the
661 :     PPC/MacOS implementation is complete, the other backends still need to
662 :     be updated.)
663 :    
664 :     ----------------------------------------------------------------------
665 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
666 : mblume 1751 Date: 2005/01/24 17:40:00 CST
667 :     Tag: blume-20050124-mlyacc
668 :     Description:
669 :    
670 :     Minor cleanup in ML-Yacc rule printing mechanism. This should fix a
671 :     problem with certain "as" patterns which previously got rendered
672 :     using incorrect syntax.
673 :    
674 :     ----------------------------------------------------------------------
675 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
676 : mblume 1750 Date: 2005/01/18 12:00:00 CST
677 :     Tag: blume-20050118-profile
678 :     Description:
679 :    
680 :     Made time profiling code (interrupt handler) in runtime system aware
681 :     of new array representation.
682 :    
683 :     ----------------------------------------------------------------------
684 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
685 : mblume 1747 Date: 2005/01/14 18:00:00 CST
686 :     Tag: blume-20050114-heap2exec
687 :     Description:
688 :    
689 :     Implemented new (but still experimental) heap2exec facility. This is
690 :     tested under Mac OS X and should work under Linux (will test shortly).
691 :     It will probably also work on the Sparc (will test some time later).
692 :    
693 :     - removed old "HACKED_STANDALONE" hack from runtime
694 :    
695 :     To be able to test this, uncomment the request for "heap2asm" in
696 :     config/targets prior to installation. (Notice that this is different
697 :     from "heap2exec" mentioned below. Not a typo.)
698 :    
699 :     To perform an actual test, run the command
700 :    
701 :     $ bin/heap2exec heapfile execfile
702 :    
703 :     (You can put heap2exec on your shell's path.)
704 :    
705 :     For example, run
706 :    
707 :     $ bin/heap2exec bin/.heap/ml-yacc.ppc-darwin mly
708 :    
709 :     This will create a standalone executable called "mly" which you can
710 :     then invoke directly as a command.
711 :    
712 :     ----------------------------------------------------------------------
713 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
714 : mblume 1746 Date: 2005/01/07 11:44:00 CST
715 :     Tag: blume-20050107-mlstring
716 :     Description:
717 :    
718 :     fixed off-by-one error in ML_STRING macro (globals.c)
719 :    
720 :     ----------------------------------------------------------------------
721 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
722 : mblume 1744 Date: 2004/12/23 18:00:00 CST
723 :     Tag: blume-20041223-santa
724 :     Description:
725 :    
726 :     Made ml-build script "smarter" (but only very little).
727 :    
728 :     ----------------------------------------------------------------------
729 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
730 : mblume 1742 Date: 2004/12/21 15:05:00 CST
731 :     Tag: blume-20041221-longlong
732 :     Description:
733 :    
734 :     * Implemented access to signed and unsigned long long data in NLFFI.
735 :     (The parameter-passing part of the picture has not complete. But
736 :     data structure access seems to work.)
737 :    
738 :     * Fixed CM's incorrect assumption that the PPC is little-endian.
739 :     (On the Mac, it is big-endian. And that's currently our only
740 :     PPC platform.)
741 :    
742 :     ----------------------------------------------------------------------
743 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
744 : mblume 1741 Date: 2004/12/21 12:50:00 CST
745 :     Tag: blume-20041221-memory
746 :     Description:
747 :    
748 :     Some cleanup in the $c/memory.cm library: separated some concerns by
749 :     moving allocation code and memory access code each into their own
750 :     files.
751 :    
752 :     ----------------------------------------------------------------------
753 : allenleung 1740 Name: Allen Leung (leunga (at) reservoir (dot) com)
754 :     Date: 2004/12/17 16:12:00 EST
755 :     Tag: leunga-20041217-cygwin-smlnj-home
756 :     Description:
757 :    
758 : blume 2155 The Unix I/O library of SML/NJ on cygwin does not understand
759 :     Windows style pathname, so problems arise when SMLNJ_HOME is set
760 :     to a Windows style pathname. _run-sml now converts SMLNJ_HOME
761 :     to a POSIX pathname on cygwin.
762 : allenleung 1740
763 :     ----------------------------------------------------------------------
764 : mblume 1737 Name: Matthias Blume (blume (at) tti - c (dot) org)
765 : mblume 1736 Date: 2004/12/16 13:35:00 CST
766 :     Tag: Release_110_52
767 :     Description:
768 :    
769 :     Last-minute changes incorporated into 110.52. Release tag moved.
770 :    
771 :     The changes:
772 :     - HashString.hashString' -> HashString.hashSubstring
773 :     - bug fix in UnivariateStats
774 :    
775 :     ----------------------------------------------------------------------
776 : mblume 1702 Name: Matthias Blume (blume (at) tti - c (dot) org)
777 : mblume 1733 Date: 2004/12/15 23:40:00 CST
778 :     Tag: blume-20041215-hashSubstring
779 :     Description:
780 :    
781 :     - HashString.hashString' -> HashString.hashSubstring
782 :     - corresponding changes in atom.sml
783 :     - "de-compressed" (aka. un-obfuscated) code for UnivariateStats and
784 :     added some comments
785 :    
786 :     ----------------------------------------------------------------------
787 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
788 : mblume 1728 Date: 2004/12/15 15:30:00 CST
789 : mblume 1736 Tag: (Release_110_52)
790 : mblume 1728 Description:
791 :    
792 :     New working version (110.52). NEW BOOTFILES!
793 :    
794 :     ----------------------------------------------------------------------
795 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
796 : mblume 1727 Date: 2004/12/15 12:45:00 CST
797 :     Tag: blume-20041215-spaces
798 :     Description:
799 :    
800 :     More on the space problem (this time for Win32).
801 :    
802 :     ----------------------------------------------------------------------
803 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
804 : mblume 1725 Date: 2004/12/14 17:30:00 CST
805 :     Tag: blume-20041214-spaces
806 :     Description:
807 :    
808 :     Hacked some of the scripts (in particular: the installer) to cope with
809 :     spaces in filenames a bit better. But beware: the current "solution"
810 :     is likely still full of bugs and inherently incomplete. (We need to
811 :     do away with those shell scripts for a comprehensive solution.)
812 :    
813 :     ----------------------------------------------------------------------
814 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
815 : mblume 1724 Date: 2004/12/13 14:45:00 CST
816 :     Tag: blume-20041213-ml-makedepend
817 :     Description:
818 :    
819 :     Fixed bug in code for ml-makedepend.
820 :    
821 :     ----------------------------------------------------------------------
822 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
823 : mblume 1720 Date: 2004/12/09 16:30:00 CST
824 :     Tag: blume-20041209-statistics
825 :     Description:
826 :    
827 :     Added two simple but potentially useful statistics modules to SML/NJ Library.
828 :     (See CHANGES file there.)
829 :    
830 :     ----------------------------------------------------------------------
831 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
832 : mblume 1718 Date: 2004/12/01 16:50:00 CST
833 :     Tag: blume-20041201-atom
834 :     Description:
835 :    
836 :     smlnj-lib:
837 :    
838 :     Added function HashString.hashString' for substrings.
839 :     Hand-inlined CharVector.foldl into HashString (for speed).
840 :     Modified implementation of structure Atom to avoid extracting
841 :     strings from substrings unless necessary.
842 :     (Also see CHANGES file for smlnj-lib.)
843 :    
844 :     ----------------------------------------------------------------------
845 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
846 : mblume 1717 Date: 2004/11/24 22:15:00 CST
847 :     Tag: blume-20041124-cml
848 :     Description:
849 :    
850 :     Made sure CML compiles when Position = Int64.
851 :    
852 :     ----------------------------------------------------------------------
853 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
854 : mblume 1715 Date: 2004/11/24 14:30:00 CST
855 :     Tag: blume-20041124-position
856 :     Description:
857 :    
858 :     The compiler can now be compiled in a mode that makes structure Position
859 :     equal to Int64. The default, however, is unchanged (Position = Int31)
860 :     for the time being.
861 :    
862 :     To enable 64-bit positions, use the following procedure:
863 :    
864 :     1. Start sml
865 :     2. Autoload $smlnj/cmb.cm (if not already autoloaded)
866 :     3. Type
867 :     #set (CMB.symval "USE_64_BIT_POSITIONS") (SOME 1);
868 :     4. Run CMB.make() as usual.
869 :    
870 :     This is barely tested. The only test so far was a little SML program
871 :     counting the number of characters in an 8-gigabyte file by
872 :     reading it character-by-character. That test was successful.
873 :    
874 :     In support of 64-bit positions, a number of new functions have been
875 :     added to the runtime system.
876 :    
877 :     ----------------------------------------------------------------------
878 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
879 : mblume 1716 Date: 2004/11/23 14:45:00 CST
880 :     Tag: blume-20041123-useFile
881 : mblume 1712 Description:
882 :    
883 :     Fixed a problem with unhelpful error messages related to problems with
884 :     .cm- or .sml files that appear as part of the sml command line.
885 :    
886 :     ----------------------------------------------------------------------
887 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
888 : mblume 1703 Date: 2004/11/18 15:40:00 CST
889 :     Tag: Release_110_51
890 :     Description:
891 :    
892 :     New working version (110.51). NEW BOOTFILES!
893 :    
894 :     ----------------------------------------------------------------------
895 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
896 : mblume 1702 Date: 2004/11/18 15:35:00 CST
897 :     Tag: <none>
898 :     Description:
899 :    
900 :     Enabled dlopen and friends for FreeBSD (as recommended by Johannes 5
901 :     Joemann).
902 :    
903 :     ----------------------------------------------------------------------
904 : allenleung 1701 Name: Allen Leung (leunga (at) reservoir (dot) com)
905 :     Date: 2004/11/17 16:05:21 EST 2004
906 :     Tag: leunga-20041117-mlrisc-live-kill
907 :     Description:
908 :    
909 : blume 2155 Added support for MLTree constructs LIVE and KILL
910 : allenleung 1701 to all the architectures.
911 :    
912 :     ----------------------------------------------------------------------
913 : mblume 1687 Name: Matthias Blume (blume (at) tti - c (dot) org)
914 : mblume 1689 Date: 2004/11/13 00:20:00 CST
915 :     Tag: blume-20041113-versiontool
916 :     Description:
917 :    
918 :     - Stripped down the versiontool: It now only handles the version number.
919 :     The date string is generated at bootstrap time (during makeml).
920 :    
921 :     - In a previous commit, fixed a minor issue with how polyequal is being
922 :     translated. In particular, the code now "looks through" abstractions.
923 :     This results in slightly fewer polyEqual warnings and hopefully slightly
924 :     more efficient code. Important examples for where this matters are
925 :     the new int64 and word64 types.
926 :    
927 :     ----------------------------------------------------------------------
928 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
929 : mblume 1687 Date: 2004/11/12 00:30:00 CST
930 :     Tag: blume-20041112-int64
931 :     Description:
932 :    
933 :     Structure Int64 fully hooked in. (The implementation is not very
934 :     efficient, though.)
935 :    
936 :     ----------------------------------------------------------------------
937 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
938 : mblume 1686 Date: 2004/11/11 17:30:00 CST
939 :     Tag: blume-20041111-more64
940 :     Description:
941 :    
942 :     All the pieces of Word64 are now there, with the exception of the
943 :     conversions from and to LargeWord. (Eventually these need to be identities,
944 :     but for the time being they don't even make sense because LargeWord is
945 :     32-bit wide.)
946 :    
947 :     Also started to add similar support for Int64, but major pieces of that
948 :     are still missing.
949 :    
950 :     ----------------------------------------------------------------------
951 : mblume 1668 Name: Matthias Blume (blume (at) tti - c (dot) org)
952 : mblume 1684 Date: 2004/11/11 00:15:00 CST
953 :     Tag: blume-20041111-word64
954 :     Description:
955 :    
956 :     Structure Word64 is now (almost) complete, word literals and patterns
957 :     seem to work. There are a few odd pieces missing. In particular,
958 :     I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
959 :     at the moment.
960 :    
961 :     Making Word64 official would mean that LargeWord becomes Word64. But
962 :     this requires extreme care because most word-word conversions have to
963 :     go through LargeWord, so making a mistake means loss of efficiency or
964 :     worse. Eventually there will be a solution similar to (but actually
965 :     simpler than) what I did with IntInf.
966 :    
967 :     ----------------------------------------------------------------------
968 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
969 : mblume 1683 Date: 2004/11/10 18:12:00 CST
970 :     Tag: blume-20041110-64bit
971 :     Description:
972 :    
973 :     More 64-bit hacking (but still not even half-way there yet).
974 :     Also, some assorted improvements to the handling of 8-bit words.
975 :    
976 :     ----------------------------------------------------------------------
977 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
978 : mblume 1682 Date: 2004/11/09 17:50:00 CST
979 : mblume 1683 Tag: <none>
980 : mblume 1682 Description:
981 :    
982 :     Started adding some infrastructure for supporting 64-bit int- and
983 :     word-types. (Still in its very early stages.)
984 :    
985 :     ----------------------------------------------------------------------
986 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
987 : mblume 1672 Date: 2004/10/28 10:45:00 CDT
988 :     Tag: Release_110_50
989 :     Description:
990 :    
991 :     New working version (110.50). NEW BOOTFILES!
992 :    
993 :     =====================
994 :    
995 :     Also:
996 :    
997 :     - Changend config/srcarchiveurl from a file just
998 :     containing the URL string into a file containing
999 :     shell script code. The code has access to the $VERSION variable.
1000 :     - Made corresponding changes to config/install.sh and config/unpack.
1001 :     - Default contents of config/srcarchiveurl uses $VERSION and
1002 :     normally does not have to be edited to reflect a version change.
1003 :    
1004 :     (As a result, a version change can be done by just editing
1005 :     config/version, the rest is now automatic.)
1006 :    
1007 :     ----------------------------------------------------------------------
1008 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1009 : mblume 1669 Date: 2004/10/27 17:50:00 CDT
1010 :     Tag: blume-20041027-btrace-msg
1011 :     Description:
1012 :    
1013 :     BackTrace.monitor now also reports the source of the exception that
1014 :     triggered the trace.
1015 :    
1016 :     ----------------------------------------------------------------------
1017 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1018 : mblume 1668 Date: 2004/10/27 17:20:00 CDT
1019 :     Tag: blume-20041027-x86-c-calls
1020 :     Description:
1021 :    
1022 :     This is the HISTORY entry for two earlier commits, both concerning
1023 :     the x86 c-calls code in MLRISC:
1024 :    
1025 :     - added a missing LOAD in the code that deals with struct arguments
1026 :     - made sure the caller does not add the wrong number of bytes to the
1027 :     stack pointer after a call of a function returning a struct
1028 :     (the callee already pops the implicit argument which points to
1029 :     the space reserved for the result)
1030 :    
1031 :     ----------------------------------------------------------------------
1032 : allenleung 1662 Name: Allen Leung (leunga (at) reservoir (dot) com)
1033 :     Date: 2004/10/24 14:00:00 EST
1034 :     Tag: leunga-20041024-x86-gas-fucomip
1035 :     Description:
1036 :    
1037 :     John discovered a bug in the syntax of fucomip.
1038 :     The opcodes FU?COMIP? have been changed to
1039 :    
1040 :     fu?comip? %st(i), %st
1041 :    
1042 :     ----------------------------------------------------------------------
1043 : mblume 1622 Name: Matthias Blume (blume (at) tti - c (dot) org)
1044 : mblume 1656 Date: 2004/10/20 15:06:00 CDT
1045 :     Tag: blume-20041020-standalone-backtrace
1046 :     Description:
1047 :    
1048 :     Added a mechanism for getting back-trace information from standalone
1049 :     programs. Here is how it works:
1050 :    
1051 :     1. The part of the program from which you want to get backtrace
1052 :     information (usually the whole program) should be wrapped with
1053 :     BackTrace.monitor. This is a (unit->'a)->'a function, and your
1054 :     main program could be modified from something like
1055 :    
1056 :     fun main (pgm, args) = ...
1057 :    
1058 :     to
1059 :    
1060 :     fun main (pgm, args) = BackTrace.monitor (fn () => ...)
1061 :    
1062 :     2. To be able to access BackTrace.monitor, you have to add
1063 :     library $smlnj-tdp/plugins.cm to the .cm file that contains your
1064 :     main function.
1065 :    
1066 :     3. Remove all compiled code (i.e., all the .cm/ subdirectories that
1067 :     CM might have created in the past for your project).
1068 :    
1069 :     4. Build the system using this command line:
1070 :    
1071 :     ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
1072 :     myprog.cm MyProg.main myprog
1073 :    
1074 :     instead of the usual
1075 :    
1076 :     ml-build myprog.cm MyProg.main myprog
1077 :    
1078 :     I changed a library name:
1079 :    
1080 :     $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
1081 :    
1082 :     New libraries:
1083 :    
1084 :     $smlnj-tdp/back-trace.cm
1085 :     -- when loaded causes the back-trace plugin to be installed
1086 :     $smlnj-tdp/coverage.cm
1087 :     -- when loaded causes the coverage plugin to be installed
1088 :    
1089 :     ----------------------------------------------------------------------
1090 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1091 : mblume 1655 Date: 2004/10/18 16:45:00 CDT
1092 :     Tag: blume-20041018-groupowner
1093 :     Description:
1094 :    
1095 :     Added an "obsolete" warning for the "group owner" syntax to CM's parser.
1096 :    
1097 :     Eliminated group owner specs from .cm files throughout the source tree.
1098 :    
1099 :     ----------------------------------------------------------------------
1100 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1101 : mblume 1653 Date: 2004/10/15 15:45:00 CDT
1102 :     Tag: blume-20041015-coverage
1103 :     Description:
1104 :    
1105 :     * Test coverage tool added!
1106 :     * Further reorganization of tracing-, debugging-, and profiling support:
1107 :    
1108 :     - moved original BTImp -- now called BackTrace -- into a separate
1109 :     library called $/trace-debug-profile.cm
1110 :     - eliminated all mentions of BTrace from SMLofNJ.Internals
1111 :     - only the instrumentation mechanism is now left in the compiler proper
1112 :     - BackTrace module is a plugin which is NOT plugged in by default
1113 :     - Coverage module is another such plugin
1114 :    
1115 :     To get the benefits of any of these plugin modules, the code in
1116 :     question must be compiled with tdp instrumentation turned on. This
1117 :     can be done by setting SMLofNJ.Internals.TDP.mode to true.
1118 :     (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
1119 :    
1120 :     Plugins are selected at link time. (Pre-compiled instrumented code
1121 :     can be re-loaded with different plugins in effect.) When an
1122 :     instrumented module is linked, whatever plugins are at that time
1123 :     enabled will come into effect for that module.
1124 :    
1125 :     To enable the back-trace plugin, load library $/trace-debug-profile.cm
1126 :     and invoke BackTrace.install() (e.g., from the interactive prompt).
1127 :     To enable the coverage plugin, load the same library and invoke
1128 :     Coverage.install().
1129 :    
1130 :     Back-traces are generated automatically on uncaught exceptions and
1131 :     when the code in question explicitly invokes BackTrace.trigger().
1132 :    
1133 :     Coverage (and execution frequency-) information must be queried
1134 :     explicitly by calling Coverage.not_covered and Coverage.hot_spots.
1135 :    
1136 :     ----------------------------------------------------------------------
1137 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1138 : mblume 1652 Date: 2004/10/14 17:40:00 CDT
1139 :     Tag: blume-20041014-tdp-core
1140 :     Description:
1141 :    
1142 :     Snapshot of a significant overhaul of how the trace/debug/profile support
1143 :     is hooked into the system (specifically: Core and SMLofNJ.Internals).
1144 :    
1145 :     ----------------------------------------------------------------------
1146 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1147 : mblume 1650 Date: 2004/10/13 16:34:00 CDT
1148 :     Tag: blume-20041013-tdp
1149 :     Description:
1150 :    
1151 :     Some rationalization of names:
1152 :    
1153 :     structure BTrace -> structure TDPInstrument
1154 :     etc.
1155 :    
1156 :     This is is preparation of using the original back-trace
1157 :     instrumentation for other purposes. "TDP" stands for
1158 :     Trace/Debug/Profile.
1159 :    
1160 :     The control flag controlling whether instrumentation is on or off is now
1161 :     registered under a different name, so instead of running sml as
1162 :    
1163 :     sml -Cinstrument.btrace-mode=true
1164 :    
1165 :     one has to say
1166 :    
1167 :     sml -Ctdp.instrument=true
1168 :    
1169 :     ----------------------------------------------------------------------
1170 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1171 : mblume 1645 Date: 2004/10/11 16:37:00 CDT
1172 :     Tag: blume-20041011-regions
1173 :     Description:
1174 :    
1175 :     Made some minor modifications to elabcore.sml to have source regions
1176 :     be propagated more tightly -- resulting in better (i.e., smaller)
1177 :     regions being reported in error- and debug messages.
1178 :    
1179 :     ----------------------------------------------------------------------
1180 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1181 : mblume 1644 Date: 2004/10/08 22:50:00 CDT
1182 :     Tag: blume-20041008-cmkw
1183 :     Description:
1184 :    
1185 :     Fixed handling of keywords in .cm files: After seeing "is" the lexer
1186 :     treats subsequent occurrences of "group", "library", "source", "is",
1187 :     "*", and "-" as ordinary identifiers rather than keywords.
1188 :    
1189 :     Most seriously, this fixes a problem with CM's "shell" tool. The tool
1190 :     is supposed to accept a tool argument called "source", but this did
1191 :     not work because of the clash with the keyword.
1192 :    
1193 :     ----------------------------------------------------------------------
1194 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1195 : mblume 1642 Date: 2004/10/07 16:00:00 CDT
1196 :     Tag: blume-20041007-cleanup
1197 :     Description:
1198 :    
1199 :     Assorted cleanup work:
1200 :    
1201 :     - got rid of intstrmap in favor of using the library's
1202 :     hash table implementation
1203 :     - threw out most of the pathnames stuff, as it was not used anyway
1204 :     - simplified tokentable implementation
1205 :     - fixed some minor spelling errors
1206 :    
1207 :     ----------------------------------------------------------------------
1208 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1209 : mblume 1641 Date: 2004/10/06 15:15:15 CDT
1210 :     Tag: blume-20041006-handler
1211 :     Description:
1212 :    
1213 :     Cleaned up the absyn to reflect the invariant that HANDLE always
1214 :     carries a FNexp as part of the type definition. This eliminates some
1215 :     superfluous sanity checks at runtime down the road.
1216 :    
1217 :     Some minor cleanup of the btrace code.
1218 :    
1219 :     ----------------------------------------------------------------------
1220 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1221 : mblume 1640 Date: 2004/10/01 10:20:30 CDT
1222 :     Tag: blume-20041001-slave
1223 :     Description:
1224 :    
1225 :     Added hack to make slave mode work in the presence of the version
1226 :     tool. (Still, since the master does two passes over the code for
1227 :     CMB.make, the release number gets bumped twice when slaves are
1228 :     attached. I don't know if this is worth fixing...)
1229 :    
1230 :     ----------------------------------------------------------------------
1231 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1232 : mblume 1639 Date: 2004/09/30 10:55:00 CDT
1233 :     Tag: blume-20040930-version
1234 :     Description:
1235 :    
1236 :     * Moved the "version" magic into its own little library under
1237 : mblume 1672 src/system/smlnj/internal. This avoids expensive reconstruction of
1238 :     a stable src/compiler/core.cm.
1239 : mblume 1639
1240 : mblume 1672 * At the same time, structure CompilerVersion is now known as
1241 :     structure SMLNJVersion.
1242 : mblume 1639
1243 :     * Arranged for the version tool to NOT kick in when rebuilding the system
1244 :     (makeml -rebuild, fixpt). Otherwise one would never reach a fixpoint.
1245 :     Also, loading the versiontool does not work when rebuilding the system
1246 :     because CM is not properly initialized at that time.
1247 :    
1248 :     ----------------------------------------------------------------------
1249 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1250 : mblume 1638 Date: 2004/09/29 14:00:00 CDT
1251 :     Tag: blume-20040929-autoversion
1252 :     Description:
1253 :    
1254 :     Implemented some CM magic to have
1255 :     file src/compiler/TopLevel/main/version.sml
1256 :     generated automagically.
1257 :     The version is taken from two files: config/version and config/release.
1258 :     The first is expected to contain a two-part version number such as 110.49.
1259 :     The second should contain a single number, but it may be missing.
1260 :    
1261 :     If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
1262 :     time the version tool is loaded (which is the first time you say CMB.make),
1263 :     then the tool will increment the value stored in config/release every
1264 :     time CMB.make is invoked.
1265 :    
1266 :     The binfile format is now insensitive to anything beyond the first
1267 :     two components of a version number, so bumping the release does not render
1268 :     binfiles incompatible. Auto-bumping can be used to keep track of versions
1269 :     during development without invalidating existing binfiles.
1270 :    
1271 :     In any case, every CMB.make updates the date information in version.sml.
1272 :     (This is the date that is printed in the banner.)
1273 :    
1274 :     ----------------------------------------------------------------------
1275 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1276 : mblume 1634 Date: 2004/09/28 10:53:00 CDT
1277 :     Tag: blume-20040928-controls
1278 :     Description:
1279 :    
1280 :     Some cleanup of the controls code.
1281 :    
1282 :     ----------------------------------------------------------------------
1283 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1284 : mblume 1633 Date: 2004/09/27 22:08:00 CDT
1285 :     Tag: blume-20040927-controls
1286 :     Description:
1287 :    
1288 :     Added two pieces of functionality to the Controls interface:
1289 :    
1290 :     1. val save'restore: 'a control -> unit -> unit
1291 :    
1292 :     grabs the current value of the control in stage 1 and restores it
1293 :     in stage 2.
1294 :    
1295 :     2. val set' : 'a control * 'a -> unit -> unit
1296 :    
1297 :     stores the given value into the control in stage 2 (i.e., delayed)
1298 :     but does all error checking in stage 1.
1299 :     (This is for string controls that need to do parse their argument --
1300 :     something that might fail. In some cases, notably in CM, one
1301 :     already knows the intended argument but wants to delay the actual
1302 :     assignment until a time when error recovery would be more difficult.)
1303 :    
1304 :     Changed the handling of controls in tool arguments to classes "sml" and
1305 :     "lazysml":
1306 :     - use Controls.save'restore as a more robust way of restoring the
1307 :     old value (in particular: without having to re-parse the string)
1308 :     - use controls to handle the "overload" keyword in the init group
1309 :     (I believe this change actually fixes a long-standing obscure bug.)
1310 :    
1311 :     ----------------------------------------------------------------------
1312 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1313 : mblume 1632 Date: 2004/09/27 17:00:00 CDT
1314 :     Tag: blume-20040927-lazysml
1315 :     Description:
1316 :    
1317 :     Added a new tool class called "lazysml" to CM's tool chest. The only
1318 :     difference to "sml" is that compilation is done with Control.lazysml
1319 :     set to true. A source of class "lazysml" is automatically recognized
1320 :     by a file name suffix of ".lml".
1321 :    
1322 :     In addition to the above feature, the original class "sml" now also
1323 :     supports a tool argument "lazy" which has the same effect. As a
1324 :     result, the following three lines are equivalent:
1325 :    
1326 :     foo.sml : lazysml
1327 :     foo.sml : sml (lazy)
1328 :     foo.sml (lazy)
1329 :    
1330 :     The setting goes into effect both during parsing and during
1331 :     compilation. The original setting is restored right after parsing and
1332 :     after compilation, respectively.
1333 :    
1334 :     In addition to all the above, there is also a general mechanism to set
1335 :     ANY of the "controls" that are available at the command line via
1336 :     "-C..." on a per-sml-file basis. The same rules that apply for "lazy"
1337 :     apply as well. (In fact, "lazy" is implemented as a special case of
1338 :     the general mechanism.)
1339 :    
1340 :     The .cm file syntax uses a new keyword tool argument called "with".
1341 :     There are several ways of indicating the desired settings:
1342 :    
1343 :     foo.sml (with:parser.quotations=true)
1344 :     foo.sml (with:(name:parser.quotations value:true))
1345 :     foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
1346 :     foo.sml (with:(name1=value1 name2=value2 ...))
1347 :     foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
1348 :    
1349 :     etc.
1350 :    
1351 :     Another possible abbreviation is to leave out the =v or value:v part
1352 :     if the name refers to a boolean control (in which case the value is
1353 :     taken to be true). Thus, one could get lazy sml also by saying:
1354 :    
1355 :     foo.sml (with:parser.lazy-keyword=true)
1356 :     foo.sml (with:parser.lazy-keyword)
1357 :     foo.sml (with:(name:parser.lazy-keyword value:true))
1358 :     foo.sml (with:(name:parser.lazy-keyword))
1359 :    
1360 :     ----------------------------------------------------------------------
1361 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1362 : mblume 1631 Date: 2004/09/24 16:22:00 CDT
1363 :     Tag: blume-20040924-ppc-long-branch
1364 :     Description:
1365 :    
1366 :     Turned message about "emiting long form of branch" off by default.
1367 :     Added a control flag to turn it back on when desired.
1368 :    
1369 :     ----------------------------------------------------------------------
1370 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1371 : mblume 1630 Date: 2004/09/24 16:05:00 CDT
1372 :     Tag: blume-20040924-rounding
1373 :     Description:
1374 :    
1375 :     Applied patch for setting rounding modes under Mac OS X. Thanks to
1376 :     Melissa O'Neill for providing the code!
1377 :    
1378 :     ----------------------------------------------------------------------
1379 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1380 : mblume 1629 Date: 2004/09/23 17:30:00 CDT
1381 :     Tag: blume-20040923-envvars
1382 :     Description:
1383 :    
1384 :     1. Changed definition of type ControlRegistry.registry_tree to
1385 :     include control_info (i.e., the name of the controlling environment
1386 :     variable).
1387 :    
1388 :     2. Added command-line flags -e and -E to print the names of environment
1389 :     variables that can be used to control internal settings.
1390 :     (This uses the new API mentioned in 1.)
1391 :    
1392 :     ----------------------------------------------------------------------
1393 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1394 : mblume 1622 Date: 2004/09/13 16:50:00 CDT
1395 :     Tag: Release_110_49
1396 :     Description:
1397 :    
1398 :     New working version (110.49). NEW BOOTFILES!
1399 :    
1400 :     ----------------------------------------------------------------------
1401 : mblume 1621 Name: Matthias Blume
1402 :     Date: 2004/09/13 16:20:00 CDT
1403 :     Tag: blume-20040913-config-mlrisc
1404 :     Description:
1405 :    
1406 :     Put target "mlrisc" back into the default list.
1407 :     (There is no harm in having it, and some users have expressed their
1408 :     wish to have "mlrisc" included by default.)
1409 :    
1410 :     ----------------------------------------------------------------------
1411 : jhr 1620 Name: John Reppy
1412 :     Date: 2004/09/13
1413 :     Tag: jhr-20040913-signals
1414 :     Description:
1415 :    
1416 :     Fixed the signal masking code to properly nest mask/unmask operations
1417 :     on a per-signal basis.
1418 :    
1419 :     ----------------------------------------------------------------------
1420 : mblume 1619 Name: Matthias Blume (blume (at) tti - c (dot) org)
1421 :     Date: 2004/09/08 13:20:00 CDT
1422 :     Tag: blume-20040908-heap-magic
1423 :     Description:
1424 :    
1425 :     Bumped the heap macig number to 0x09082004 to account for the changed
1426 :     layout of the ML frame under MacOS X.
1427 :    
1428 :     ----------------------------------------------------------------------
1429 : allenleung 1618 Name: Allen Leung (leunga (at) reservoir (dot) com)
1430 :     Date: 2004/09/03 11:26:00 EST
1431 :     Tag: leunga-20040903-cygwin-install
1432 :     Description:
1433 :    
1434 :     Added a patch to _arch-n-opsys to enable the Cygwin runtime.
1435 :     The Cygwin runtime is turned on by setting the environment
1436 :     variable SMLNJ_CYGWIN_RUNTIME to 1.
1437 :    
1438 :     ----------------------------------------------------------------------
1439 : mblume 1615 Name: Matthias Blume (blume (at) tti - c (dot) org)
1440 : mblume 1617 Date: 2004/08/31 17:15:00 CDT
1441 :     Tag: blume-20040831-core
1442 :     Description:
1443 :    
1444 :     Added some exports to src/compiler/core.cm upon request by J. Joemann.
1445 :    
1446 :     ----------------------------------------------------------------------
1447 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1448 : mblume 1616 Date: 2004/08/30 17:55:00 CDT
1449 :     Tag: blume-20040830-installer
1450 :     Description:
1451 :    
1452 :     Upon request by Johannes Joemann:
1453 :    
1454 :     - improved ML code of installer to fall back to coping when renaming
1455 :     fails (i.e., when source and target are on different file systems);
1456 :     the code compiles but has yet to be tested in anger
1457 :     - removed mlrisc from list of default targets (config/targets)
1458 :    
1459 :     ----------------------------------------------------------------------
1460 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1461 : mblume 1615 Date: 2004/08/27 17:20:00 CDT
1462 :     Tag: blume-20040827-ptreql
1463 :     Description:
1464 :    
1465 :     Added ptreql primop to structure InlineT (upon request from Larry
1466 :     Paulson).
1467 :    
1468 :     ----------------------------------------------------------------------
1469 : allenleung 1613 Name: Allen Leung (leunga (at) reservoir (dot) com)
1470 : allenleung 1614 Date: 2004/08/15 21:21:00 EST
1471 :     Tag: leunga-110_48-udgraph
1472 :     Description:
1473 :    
1474 :     Another bug fix from Carl Hauser:
1475 :    
1476 :     diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
1477 :     > 48c48
1478 :     > < | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
1479 :     > e::L)
1480 :     > ---
1481 :     > > | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
1482 :     > Without this, any deletion of an edge in an undirected graph does severe
1483 :     > violence to the graph.
1484 :    
1485 :     ----------------------------------------------------------------------
1486 :     Name: Allen Leung (leunga (at) reservoir (dot) com)
1487 : allenleung 1613 Date: 2004/08/10 23:35:00 EST
1488 :     Tag: leunga-110_48-ppc
1489 :     Description:
1490 :    
1491 :     The IBM/MacOS syntax switch on PPC was incorrectly swapped.
1492 :    
1493 :     ----------------------------------------------------------------------
1494 : mblume 1602 Name: Matthias Blume (blume (at) tti - c (dot) org)
1495 : mblume 1615 Date: 2004/08/10 12:00:00 CDT
1496 : mblume 1602 Tag: Release_110_48
1497 :     Description:
1498 :    
1499 :     New working version (110.48). NEW BOOTFILES!
1500 :    
1501 :     ----------------------------------------------------------------------
1502 : allenleung 1599 Name: Allen Leung (leunga (at) reservoir (dot) com)
1503 :     Date: 2004/08/09 12:21:00 EST
1504 :     Tag: leunga-110_47-dijsktra
1505 :     Description:
1506 :    
1507 :     Bug fix from Carl Hauser:
1508 :    
1509 : blume 2155 single_source_shortest_paths in dijkstra.sml was observed to get wrong
1510 : allenleung 1599 answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
1511 :    
1512 : blume 2155 The problem is that following the expression A.update(dist,s,Num.zero)
1513 : allenleung 1599 it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
1514 :    
1515 :     ----------------------------------------------------------------------
1516 : mblume 1595 Name: Matthias Blume (blume (at) tti - c (dot) org)
1517 :     Date: 2004/08/06 18:10:00 CDT
1518 :     Tag: blume-20040806-cmdline
1519 :     Description:
1520 :    
1521 :     Fiddled with handling of command-line options:
1522 :    
1523 :     * sml now quits after processing the command line
1524 :     if -H, -S, -h<n>, or -s<n> appears as the last
1525 :     command-line argument
1526 :     * a new option -q terminates the session when encountered on
1527 :     the command line; subsequent arguments will be ignored
1528 :     * bug fixes: short (erroneous) arguments are no longer ignored
1529 :     completely
1530 :    
1531 :     ----------------------------------------------------------------------
1532 : allenleung 1593 Name: Allen Leung (leunga (at) reservoir (dot) com)
1533 :     Date: 2004/08/04 18:17:00 EST
1534 :     Tag: leunga-110_47-ppc-ibm-asm
1535 :     Description:
1536 :    
1537 :     - Added minimal IBM assembly syntax support for PowerPC.
1538 :    
1539 : blume 2155 - Cygwin: manually changed the file cygwin.def. Some exported symbols have
1540 : allenleung 1593 been altered in the runtime. We need an automatic way to keep the file
1541 :     in sync.
1542 :    
1543 :     ----------------------------------------------------------------------
1544 : mblume 1548 Name: Matthias Blume (blume (at) tti - c (dot) org)
1545 : mblume 1581 Date: 2004/08/04 14:00:00 CDT
1546 :     Tag: Release_110_47
1547 :     Description:
1548 :    
1549 :     New working version (110.47). NEW BOOTFILES!
1550 :    
1551 :     ----------------------------------------------------------------------
1552 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1553 : mblume 1580 Date: 2004/08/03 14:25:00 CDT
1554 :     Tag: blume-20040803-callingconv
1555 :     Description:
1556 :    
1557 :     Added low-level support for choosing C calling conventions by
1558 :     twiddling the type of rawccall. (See
1559 :     src/compiler/Semant/types/cproto.sml for details.)
1560 :    
1561 :     ----------------------------------------------------------------------
1562 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1563 : mblume 1579 Date: 2004/08/02 15:55:00 CDT
1564 :     Tag: blume-20040802-backout
1565 :     Description:
1566 :    
1567 :     Backed out of change to win32-filesys.c. The earlier patch to
1568 :     get_file_time caused CM to produce files with the wrong time stamp.
1569 :    
1570 :     ----------------------------------------------------------------------
1571 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1572 : mblume 1573 Date: 2004/08/02 14:45:00 CDT
1573 :     Tag: blume-20040802-nlffi-win32
1574 :     Description:
1575 :    
1576 :     Added NLFFI support for Win32, adapted from a patch provided by David
1577 :     Hansel. This is currently completely untested. Also, the issue
1578 :     concerning stdcall vs. ccall is still unresolved.
1579 :    
1580 :     ----------------------------------------------------------------------
1581 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1582 : mblume 1572 Date: 2004/07/30 17:55:00 CDT
1583 :     Tag: blume-20040730-various
1584 :     Description:
1585 :    
1586 :     Gearing up towards 110.47...
1587 :    
1588 :     - various minor bugfixes to ml-nlffigen
1589 :     - a beginning of a manual for nlffi
1590 :    
1591 :     - eliminated 'export name=value' in config/install.sh as this does
1592 :     not work with certain versions of /bin/sh
1593 :     (Thanks to David King at Motorola for catching this.)
1594 :    
1595 :     - several bugfixes provided or suggested by David Hansel at Reactive Systems:
1596 :     - added a test for tm==NULL to gmtime.c and localtime.c
1597 :     - applied patch for incorrect GetFileTime under win32
1598 :     - toSeconds -> toMilliseconds in Win32/win32-process.sml
1599 :    
1600 :     ----------------------------------------------------------------------
1601 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1602 : mblume 1557 Date: 2004/07/21 18:20:00 CDT
1603 :     Tag: blume-20040721-nlffigen
1604 :     Description:
1605 :    
1606 :     - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
1607 :     for a typedef to an incomplete type, but leave out the "typ" member.
1608 :     (This is just for consistency.)
1609 :     - Started to produce what is supposed to become better (i.e., comprehensive)
1610 :     documentation of what ml-nlffigen does and produces.
1611 :    
1612 :     ----------------------------------------------------------------------
1613 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1614 : mblume 1548 Date: 2004/07/14 16:25:00 CDT
1615 :     Tag: blume-20040714-union
1616 :     Description:
1617 :    
1618 :     Added C_UNION to c-calls/c-types.sml and updated the machinery
1619 :     (ml-nlffigen, cproto.sml) that conveys C function interface
1620 :     information to the code generator.
1621 :    
1622 :     However, the actual architecture-specific implementation of function
1623 :     arguments and results that are C unions is still not implemented.
1624 :    
1625 :     ----------------------------------------------------------------------
1626 : allenleung 1539 Name: Allen Leung (leunga (at) reservoir (dot) com)
1627 : allenleung 1546 Date: 2004/07/14 14:38:00 EST
1628 :     Tag: leunga-110_46_1-ppc-lwzu
1629 :     Description:
1630 :    
1631 :     Added these instructions to the PowerPC architecture:
1632 :     LBZU(X), LHZU(X), LWZU(X),
1633 : blume 2155 STWU(X), STFDU, STFSU
1634 :    
1635 : allenleung 1546 etc...
1636 :    
1637 :     Note: I haven't added their instruction encoding into the description.
1638 :    
1639 :     ----------------------------------------------------------------------
1640 :     Name: Allen Leung (leunga (at) reservoir (dot) com)
1641 : allenleung 1539 Date: 2004/07/13 15:04:00 EST
1642 :     Tag: leunga-110_46_1-ppc-lwarx
1643 :     Description:
1644 :    
1645 :     Added the two instructions LWARX and STWCX to the PowerPC
1646 : blume 2155 instruction set.
1647 : allenleung 1539
1648 :     A (untested) rewrite of loop-structure.sml. The old version
1649 :     is completely broken.
1650 :    
1651 :     ----------------------------------------------------------------------
1652 : mblume 1479 Name: Matthias Blume (blume (at) tti - c (dot) org)
1653 : mblume 1537 Date: 2004/07/13 13:50:00 CDT
1654 :     Tag: blume-20040713-nlffi
1655 :     Description:
1656 :    
1657 :     - use paramAlloc to report c-calls with too many arguments
1658 :     (for PPC version where parameter area is pre-allocated)
1659 :     - added ccall_maxargspace to machspec (to implement the above)
1660 :     - made "make" commend in CM's "make" tool configurable
1661 :     - added option (default: on) for passing the name of the SML/NJ's "bin"
1662 :     directory to "make"; the call looks like this:
1663 :    
1664 :     make <options> SMLNJ_BINDIR=<dir> <target>
1665 :    
1666 :     This can be used by the Makefile to, e.g., pick the "right" version
1667 :     of ml-nlffigen.
1668 :     - minor code tweaks
1669 :    
1670 :     ----------------------------------------------------------------------
1671 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1672 : mblume 1532 Date: 2004/07/12 22:50:00 CDT
1673 :     Tag: blume-110_46_1-macosx-nlffi
1674 :     Description:
1675 :    
1676 :     NLFFI under Mac OS X now working (sort of). This is largely untested,
1677 :     though.
1678 :    
1679 :     Note: 1. You have to make a new, clean build of the runtime system.
1680 :     2. There are new BOOTFILES, you have to use them!
1681 :     (Doing the bootstrap process yourself would be *very* painful!
1682 :     If you absolutely have to do it, build the system under
1683 :     a different architecture and then cross-compile.)
1684 :    
1685 :     Version bumped to 110.46.1 to account for runtime data format changes.
1686 :    
1687 :     ----------------------------------------------------------------------
1688 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1689 : mblume 1515 Date: 2004/06/18 14:30:00 CDT
1690 :     Tag: blume-20040618-unix
1691 :     Description:
1692 :    
1693 :     Changed the implementation of structure Unix so that the same stream
1694 :     is returned every time one of the {text,bin}{In,Out}streamOf functions
1695 :     is invoked on the same proc. This is not what the spec currently
1696 :     says -- although IMO it arguably should. (See discussion below.)
1697 :    
1698 :     ----------------------------------------------------------------------
1699 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1700 : mblume 1507 Date: 2004/06/17 18:15:00 CDT
1701 :     Tag: Release_110_46
1702 :     Description:
1703 :    
1704 :     New working version (110.46). NEW BOOTFILES!
1705 :    
1706 :     ----------------------------------------------------------------------
1707 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1708 : mblume 1506 Date: 2004/06/17 17:20:00 CDT
1709 :     Tag: blume-20040617-timer-unix
1710 :     Description:
1711 :    
1712 :     Changed the interface of structures Timer and Unix to match the most
1713 :     recent Basis spec.
1714 :    
1715 :     In the case of Unix there still seems to be an open/weird issue:
1716 :    
1717 :     The {text,bin}{In,Out}streamOf functions are supposed to create
1718 :     fresh streams whenever they are called -- as opposed to have them
1719 :     return the same stream every time. This design is supposed to
1720 :     prevent space leaks caused by proc values hanging on to streams.
1721 :    
1722 :     The reap function, on the other hand, is supposed to close the
1723 :     streams. This cannot be done without having a handle on the
1724 :     stream in proc after all...
1725 :    
1726 :     I took the liberty to implement the following stopgap solution:
1727 :    
1728 :     The proc value hangs on to the most recently created stream(s).
1729 :     Reap closes those. If either or both of the two streams hadn't
1730 :     been created at all yet, then reap will close the corresponding
1731 :     file descriptors directly.
1732 :    
1733 :     PS: I don't understand the original space leak argument anymore. If
1734 :     a proc hangs on to the imperative stream, then I/O operations on those
1735 :     will advance the state of the cached stream and avoid the space leak.
1736 :    
1737 :     ----------------------------------------------------------------------
1738 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1739 : mblume 1500 Date: 2004/05/28 16:45:00 CDT
1740 :     Tag: blume-20040528-basis
1741 :     Description:
1742 :    
1743 :     Added signature PACK_REAL and exported functor PrimIO.
1744 :    
1745 :     ----------------------------------------------------------------------
1746 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1747 : mblume 1498 Date: 2004/05/25 16:00:00 CDT
1748 :     Tag: blume-20040525-group-owner
1749 :     Description:
1750 :    
1751 :     CM now ignores (but still accepts) the "owner" information in group
1752 :     descriptions. The owner of a group is its next enclosing
1753 :     library. Each group must have a unique owner. (There is a virtual
1754 : mblume 1505 "toplevel" library that owns groups which are not nested within a real
1755 : mblume 1498 library.) Previously, each group had to explicitly declare its owner,
1756 :     and CM would check that such a declaration is correct. The new scheme
1757 :     is to have CM check that for each group there is precisely one owning
1758 :     library.
1759 :    
1760 :     The advantage of the new scheme is that the programmer no longer needs
1761 :     to maintain the somewhat annoying owner information. The downside is
1762 :     that CM cannot enforce the ownership rule across multiple runs of
1763 :     CM.make. Fortunately, enclosing the same group in two different
1764 :     libraries A and B which are not part of the same program does not
1765 :     cause real problems.
1766 :    
1767 :     ----------------------------------------------------------------------
1768 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1769 : mblume 1496 Date: 2004/05/20 16:00:00 CDT
1770 :     Tag: blume-20040520-win32
1771 :     Description:
1772 :    
1773 :     Made the win32 version work again. (Strangely, a misplaced comma had
1774 :     slipped into win32-process.c which prevented the runtime from being
1775 :     compiled correctly.)
1776 :    
1777 :     Also, included a minor addition to ml-build.bat analogous to what was
1778 :     done in blume-20040519-ml-build.
1779 :    
1780 :     ----------------------------------------------------------------------
1781 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1782 : mblume 1494 Date: 2004/05/19 22:10:00 CDT
1783 :     Tag: blume-20040519-ml-build
1784 :     Description:
1785 :    
1786 :     Arranged for ml-build to clean up after itself a little bit better.
1787 :     The script generates a temporary SML source file and compiles it using
1788 :     CM, so CM generates metadata (GUID, SKEL, objectfile) for it. It now
1789 :     gets rid of those at the end, so they don't accumulate under .cm.
1790 :    
1791 :     This required a minor change to install.sh because the name of the
1792 :     metadata directory (default: .cm) is actually configurable at
1793 :     installation time.
1794 :    
1795 :     ----------------------------------------------------------------------
1796 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1797 : mblume 1491 Date: 2004/05/18 15:50:00 CDT
1798 :     Tag: blume-20040518-mkreader
1799 :     Description:
1800 :    
1801 :     Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
1802 :     implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
1803 :    
1804 :     ----------------------------------------------------------------------
1805 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1806 : mblume 1489 Date: 2004/05/11 14:35:00 CDT
1807 :     Tag: blume-20040511-win32sock
1808 :     Description:
1809 :    
1810 :     Added previously missing support for many socket-related functions
1811 :     under win32. Thanks to David Hansel <hansel@reactive-systems.com>
1812 :     for the voluminous patch!
1813 :    
1814 :     (I have not tested this patch under win32 yet.)
1815 :    
1816 : mblume 1490 Here is David's e-mail:
1817 :    
1818 :     Hi,
1819 :    
1820 :     Attached to this email you find a diff against sml/nj 110.45
1821 :     that will enable socket support under Windows.
1822 :    
1823 :     To apply the patch (using unix or cygwin)
1824 :     1) gunzip runtime.diff.gz
1825 :     2) "cd" into "src/runtime" in the source tree of a fresh
1826 :     110.45 installation.
1827 :     3) patch -p 1 < [your/path/to]runtime.diff
1828 :    
1829 : blume 2155 The code compiles fine but has NOT yet been extensively tested.
1830 :     I only ran a few tests for basic socket client functionality
1831 :     (which worked fine). Especially the functions that use ioctl
1832 : mblume 1490 are not tested at all and might not work (see below).
1833 :    
1834 :     I implemented this since we want to move to a newer version of sml/nj
1835 :     but need socket support in order to use it. This is the first time I
1836 :     even had a look at the sml/nj source, so please review my changes
1837 :     before making this part of the distribution! Here are a few issues
1838 : blume 2155 that I think might be better for someone to solve who is more
1839 : mblume 1490 familiar with the sml/nj source (and socket programming):
1840 :    
1841 :     - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
1842 :     exception since I could not figure out what the windows equivalent
1843 :     of these functions is
1844 :    
1845 :     - In sockets-osdep.h there are a some #include statements that are
1846 :     only used in a few files that include sockets-osdep.h
1847 :    
1848 :     - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
1849 :     does not process its return value since I don't know how to
1850 :     report an error upwards.
1851 :    
1852 :     - It would probably be good to have a call to WSACleanup() when
1853 : blume 2155 the library is unloaded (if there is such a possibility).
1854 : mblume 1490 Otherwise I think Windows will take care of this automatically
1855 :     when the process finishes.
1856 :    
1857 :     - I used ioctlsocket() as a replacement for ioctl() but I have
1858 :     no idea if that is actually the proper replacement on Windows.
1859 :    
1860 :     - All these issues are marked in the code by "FIXME" comments.
1861 :    
1862 : blume 2155 We use sml/nj extensively in our products and are quite happy
1863 :     with it. I hope this contribution will help you.
1864 : mblume 1490
1865 :     Keep up the good work!
1866 :    
1867 :     David
1868 :    
1869 : mblume 1489 ----------------------------------------------------------------------
1870 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1871 : mblume 1488 Date: 2004/05/11 14:20:00 CDT
1872 :     Tag: blume-20040511-installml
1873 :     Description:
1874 :    
1875 :     Fixed two bugs in installml script. (Thanks to Vesa A. Norrman for
1876 :     the patch.)
1877 :    
1878 :     ----------------------------------------------------------------------
1879 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1880 : mblume 1487 Date: 2004/05/11 14:05:00 CDT
1881 :     Tag: blume-20040511-nlffi-netbsd
1882 :     Description:
1883 :    
1884 :     Added support for nlffi under netbsd. (Thanks to Vesa A. Norrman for
1885 :     the patch.)
1886 :    
1887 :     ----------------------------------------------------------------------
1888 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1889 : mblume 1486 Date: 2004/05/11 12:05:00 CDT
1890 :     Tag: blume-20040511-exports
1891 :     Description:
1892 :    
1893 :     As per request by Adam Chlipala <adam@hcoop.net>, extended various
1894 :     export lists in compiler-related .cm-files.
1895 :    
1896 :     ----------------------------------------------------------------------
1897 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1898 : mblume 1485 Date: 2004/05/11 11:35:00 CDT
1899 :     Tag: blume-20040511-allsource
1900 :     Description:
1901 :    
1902 :     The installer now honors the "src-smlnj" target again, although its meaning
1903 :     has changed from "all sources required for the compiler" to "all sources
1904 :     the installer knows about". In other words, if you enable "src-smlnj"
1905 :     in the "targets" file, then the installer will pull in sources for
1906 :     everything. (Notice that this refers to source code only. Compiled
1907 :     code is still only installed for modules that were requested explicitly
1908 :     or which are required for other modules that were requested explicitly.)
1909 :    
1910 :     ----------------------------------------------------------------------
1911 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1912 : mblume 1482 Date: 2004/04/23 17:40:00 CDT
1913 :     Tag: blume-20040423-ieee-scan
1914 :     Description:
1915 :    
1916 :     Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
1917 :     in the exponent calculation we get INF or ZERO (depending on the mantissa
1918 :     and the sign of the exponent).
1919 :    
1920 :     ----------------------------------------------------------------------
1921 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1922 : mblume 1480 Date: 2004/04/23 10:40:00 CDT
1923 :     Tag: blume-20040423-ml-build
1924 :     Description:
1925 :    
1926 :     The ml-build script now terminates with a non-0 status when something
1927 :     goes wrong.
1928 :    
1929 :     ----------------------------------------------------------------------
1930 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1931 : mblume 1479 Date: 2004/04/22 16:35:00 CDT
1932 :     Tag: blume-20040422-Option
1933 :     Description:
1934 :    
1935 :     Made exception Option to be the same as exception Option.Option
1936 :     (as it should be).
1937 :    
1938 :     ----------------------------------------------------------------------
1939 : allenleung 1468 Name: Allen Leung (leunga (at) reservoir (dot) com)
1940 :     Date: 2004/03/19 14:40:00 EST
1941 :     Tag: leunga-20040319-cygwin-nlffi
1942 :     Description:
1943 :    
1944 :     Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1945 :     of SML/NJ. The problem is that
1946 :    
1947 :     lib = dlopen(NULL, ...)
1948 :     f = dlsym(lib, "malloc");
1949 :    
1950 : blume 2155 does not work on Windows unless we explicitly export symbols
1951 :     such as 'malloc' during linking. We fixed this by explicitly
1952 : allenleung 1468 exporting the required symbols with the magic gcc incantation:
1953 :    
1954 :     -Wl,--export-all cygwin.def
1955 :    
1956 :     where cygwin.def is a file containing all the symbols that
1957 : blume 2155 we wish to export.
1958 : allenleung 1468
1959 : blume 2155 I suspect this is a Windows problem and we'll have to
1960 :     do the same (somehow with windows compilers) when
1961 : allenleung 1468 we build the native win32 version with the system
1962 : blume 2155 calls LoadLibrary/GetProcAddress.
1963 : allenleung 1468
1964 :     ----------------------------------------------------------------------
1965 : mblume 1384 Name: Matthias Blume (blume (at) tti - c (dot) org)
1966 : mblume 1465 Date: 2004/03/04 16:35:00 CST
1967 :     Tag: blume-20040304-intinf-fmt
1968 :     Description:
1969 :    
1970 :     Fixed problem with IntInf.fmt (sign would show up on the right instead
1971 :     of on the left for BIN, OCT, and HEX).
1972 :    
1973 :     ----------------------------------------------------------------------
1974 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1975 : mblume 1464 Date: 2004/03/04 11:25:00 CST
1976 :     Tag: blume-20040304-symlinks
1977 :     Description:
1978 :    
1979 :     Fixed problem with installer script (unix only) where bin/ml-yacc and
1980 :     friends pointed (via symlinks) to absolute locations instead of just
1981 :     .run-sml. This was reported by Vesa A Norrman.
1982 :    
1983 :     ----------------------------------------------------------------------
1984 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1985 : mblume 1455 Date: 2004/02/13 14:50:00 CST
1986 :     Tag: Release_110_45
1987 :     Description:
1988 :    
1989 :     New working version (110.45). New bootfiles.
1990 :    
1991 :     ----------------------------------------------------------------------
1992 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
1993 : mblume 1448 Date: 2004/01/26 15:15:15 CST
1994 :     Tag: blume-20040126-toplevel
1995 :     Description:
1996 :    
1997 :     Improved handling of exceptions at the interactive toplevel.
1998 :    
1999 :     ----------------------------------------------------------------------
2000 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2001 :     Date: 2004/01/26 11:25:00
2002 : mblume 1446 Tag: blume-20040126-app
2003 :     Description:
2004 :    
2005 :     Type of top-level "app" corrected.
2006 : mblume 1447 Added code for setting vp_limitPtrMask to Win32-specific runtime.
2007 : mblume 1446
2008 :     ----------------------------------------------------------------------
2009 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2010 : mblume 1427 Date: 2003/11/18 17:10 CST
2011 :     Tag: blume-20031118-basis-fiddle
2012 :     Description:
2013 :    
2014 :     - changed Timer interface to what might become the spec
2015 :     - POSIX_FLAGS -> BIT_FLAGS according to spec
2016 :     - some other minor discrepancies wrt. spec eliminated
2017 :    
2018 :     ----------------------------------------------------------------------
2019 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2020 : mblume 1414 Date: 2003/11/06 12:00:00 CST
2021 :     Tag: Release_110_44
2022 :     Description:
2023 :    
2024 :     New working version (110.44). New bootfiles.
2025 :    
2026 :     ----------------------------------------------------------------------
2027 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2028 : mblume 1411 Date: 2003/11/04 11:50:00 CST
2029 :     Tag: blume-20031104-move-libraries
2030 :     Description:
2031 :    
2032 :     Eliminated the "dont_move_libraries" directive in config/targets.
2033 :     (The mechanism was broken and could not be fixed easily. Moreover,
2034 :     there does not seem to be any reason not to move all libraries into
2035 :     lib during installation. I originally implemented this directive as a
2036 :     backward-compatibility feature when I first introduced the new CM.
2037 :     Now that things have been stable for a long time and going back to the
2038 :     old CM is not an option, there is no reason to keep it around.)
2039 :    
2040 :     ----------------------------------------------------------------------
2041 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2042 : mblume 1409 Date: 2003/11/03 16:00:00 CST
2043 :     Tag: blume-20031103-installdir
2044 :     Description:
2045 :    
2046 :     Made installer honor INSTALLDIR variable again. (Thanks to Chris
2047 :     Richards for pointing out the problem and providing the solution.)
2048 :    
2049 :     ----------------------------------------------------------------------
2050 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2051 : mblume 1400 Date: 2003/10/01 17:05:00 CDT
2052 :     Tag: blume-20031001-lal-mlrisc
2053 :     Description:
2054 :    
2055 :     MLRISC bug fix from Lal.
2056 :    
2057 :     ----------------------------------------------------------------------
2058 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2059 : mblume 1399 Date: 2003/09/30 16:10:00 CDT
2060 :     Tag: blume-20030930-primio-bat
2061 : mblume 1398 Description:
2062 :    
2063 :     1. Added openVector, nullRd, and nullWr to PRIM_IO.
2064 :     2. Improved .bat files (for Win32 port) to make things work under Win95.
2065 :     (thanks to Aaron S. Hawley for this one)
2066 :    
2067 :     ----------------------------------------------------------------------
2068 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2069 : mblume 1395 Date: 2003/09/26 16:05:00 CDT
2070 :     Tag: blume-20030926-wrappriv
2071 :     Description:
2072 :    
2073 :     Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
2074 :    
2075 :     ----------------------------------------------------------------------
2076 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2077 : mblume 1394 Date: 2003/09/26 15:00:00 CDT
2078 :     Tag: blume-20030926-110_43_3
2079 :     Description:
2080 :    
2081 :     - additional cleanup
2082 :     - version number bump, NEW BOOTFILES
2083 :    
2084 :     ----------------------------------------------------------------------
2085 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2086 : mblume 1393 Date: 2003/09/26 12:00:00 CDT
2087 :     Tag: blume-20030926-ppautoload
2088 :     Description:
2089 :    
2090 :     I modified the read-eval-print loop so that the autoloader gets
2091 :     invoked whenever the prettyprinter tries to look up a symbol that
2092 :     is not currently defined in the toplevel environment but which
2093 :     appears in CM's autoload registry. As a result, we see far fewer of
2094 :     those ?.Foo.Bar.xxx names in the prettyprinter's output.
2095 :    
2096 :     In addition to this I tried to clean up some pieces of the Basis
2097 :     implementation (e.g., Socket, Word8Array) in order to prevent other
2098 :     instances of these ?.Foo.Bar.xxx names from being printed.
2099 :    
2100 :     The mechanism that picks names for types still needs some work, though.
2101 :     (Right now it seems that if there is a type A.t which is defined to
2102 : mblume 1394 be B.u, but B is unavailable at toplevel, then A.t gets printed as
2103 :     "?.B.u" although the perhaps more sensible solution would be to use
2104 : mblume 1393 "A.t" in this case. In other words, the prettyprinter should follow
2105 :     a chain of DEFtycs not farther than there are corresponding toplevel
2106 :     names in the current environment.)
2107 :    
2108 :     ----------------------------------------------------------------------
2109 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2110 : mblume 1391 Date: 2003/09/24 16:31:00 CDT
2111 :     Tag: blume-20030924-installer
2112 :     Description:
2113 :    
2114 :     Another installer tweak: All the ML code for the installer is now
2115 :     compiled during CMB.make and put into a little library called
2116 :     $smlnj/installer.cm. The installation then simply invokes
2117 :    
2118 :     sml -m $smlnj/installer.cm
2119 :    
2120 :     and everything happens automagically.
2121 :    
2122 :     Win32: ML code senses value of environment variable SMLNJ_HOME.
2123 :     Unix: ML code senses values of environment variables ROOT, CONFIGDIR,
2124 :     and BINDIR.
2125 :    
2126 :     The new scheme guarantees that the ML code responsible for the installation
2127 :     is in sync with the APIs of the main system. Also, the installer is
2128 :     somewhat faster because the installer script is precompiled.
2129 :    
2130 :     ----------------------------------------------------------------------
2131 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2132 : mblume 1390 Date: 2003/09/24 15:35:00 CDT
2133 :     Tag: blume-20030924-synsock
2134 :     Description:
2135 :    
2136 :     Added a signature SYNCHRONOUS_SOCKET to basis.cm. This is like SOCKET
2137 :     but excludes all non-blocking operations. Defined SOCKET (in Basis)
2138 :     and CML_SOCKET in terms of SYNCHRONOUS_SOCKET. Removed superfluous
2139 :     implementations of non-blocking operations from CML's Socket
2140 :     structure.
2141 :    
2142 :     ----------------------------------------------------------------------
2143 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2144 : mblume 1389 Date: 2003/09/24 15:10:05 CDT
2145 :     Tag: blume-20030924-sockets
2146 :     Description:
2147 :    
2148 :     1. Fixed SOCKET API and implementation to match Basis spec.
2149 :     This required changing the internal representation of sockets to one
2150 :     that remembers (for each socket file descriptor) whether it is currently
2151 :     blocking or non-blocking. This state is maintained lazily (i.e., a system
2152 :     call is made only if the state actually needs to change).
2153 :    
2154 :     2. OS-specific details of sockets were moved into separate files, thus
2155 :     making it possible to unify the bulk of the socket implementations
2156 :     between Unix and Win32.
2157 :    
2158 :     3. CML's socket API changed accordingly.
2159 :     (Note that we need to remove non-blocking functions from this API
2160 :     since they are redundant in the case of CML!)
2161 :    
2162 :     4. CML's socket implementation now makes use of non-blocking functions
2163 :     provided by Basis, thus removing all OS-dependent code from this part
2164 :     of CML.
2165 :    
2166 :     5. Changed Real64.precision from 52 to 53. Minor cleanup in Real64 code.
2167 :    
2168 :     ----------------------------------------------------------------------
2169 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2170 : mblume 1387 Date: 2003/09/22 12:10:00 CDT
2171 : mblume 1388 Tag: blume-20030922-110_43_2
2172 : mblume 1387 Description:
2173 :    
2174 :     Made a new interim version and bootfiles for developer's bootstrapping
2175 :     convenience.
2176 :    
2177 :     110.43.2 -- NEW BOOTFILES
2178 :    
2179 :     ----------------------------------------------------------------------
2180 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2181 : mblume 1385 Date: 2003/09/19 15:55:00 CDT
2182 :     Tag: blume-20030919-cmdir
2183 :     Description:
2184 :    
2185 :     1. new-install.sh -> install.sh
2186 :     2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
2187 :     3. tweaked installer so that another name instead of .cm can be chosen
2188 :     at install time (by setting the CM_DIR_ARC environment variable
2189 :     during installation); once installation is complete, the name is
2190 :     fixed
2191 :    
2192 :     ----------------------------------------------------------------------
2193 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2194 : mblume 1384 Date: 2003/09/18 16:00:00 CDT
2195 :     Tag: blume-20030918-110_43_1
2196 :     Description:
2197 :    
2198 :     Made a new interim version and bootfiles for developer's bootstrapping
2199 :     convenience.
2200 :    
2201 :     110.43.1 -- NEW BOOTFILES
2202 :    
2203 :     ----------------------------------------------------------------------
2204 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2205 : mblume 1383 Date: 2003/09/18 15:20:00 CDT
2206 :     Tag: blume-20030918-misc
2207 :     Description:
2208 :    
2209 :     1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
2210 :     this seems to be controversial at the moment) and used those in
2211 :     Posix.ProcEnv.times.
2212 :    
2213 :     2. Added Time.{from,to}Nanoseconds to Time.
2214 :    
2215 :     3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
2216 :     For example, fromLargeInt never needs to look at more than 3 "big
2217 :     digits" to get its 53 bits of precision.
2218 :    
2219 :     ----------------------------------------------------------------------
2220 : mblume 1384 Name: Matthias Blume (blume (at) tti - c (dot) org)
2221 : mblume 1381 Date: 2003/09/17 16:30:00 CDT
2222 :     Tag: blume-20030917-real32-slices
2223 :     Description:
2224 :    
2225 :     Added an entry to the primitive environment
2226 :     (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
2227 :     added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
2228 :    
2229 :     Removed some of the "magic" constants in real64.sml and replaced them
2230 :     with code that generates these values from their corresponding
2231 :     integer counterparts.
2232 :    
2233 :     Made all(?) the slice-related changes to the Basis and made everything
2234 :     compile again...
2235 :    
2236 :     ----------------------------------------------------------------------
2237 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org)
2238 : mblume 1375 Date: 2003/09/15 17:45:00 CDT
2239 :     Tag: blume-20030915-rbase
2240 :     Description:
2241 :    
2242 :     Fixed bug in Real.fromLargeInt.
2243 :    
2244 :     ----------------------------------------------------------------------
2245 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2246 : mblume 1369 Date: 2003/09/13 18:11:00 CDT
2247 :     Tag: blume-20030913-libinstall
2248 :     Description:
2249 :    
2250 :     Minor bugfix in config/libinstall (set anchor with path to
2251 :     standalone tool after installing it, otherwise libraries that
2252 :     need ml-lex or ml-yacc won't compile the first time the installer
2253 :     runs).
2254 :    
2255 :     ----------------------------------------------------------------------
2256 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2257 : mblume 1368 Date: 2003/09/12 11:45:00 CDT
2258 :     Tag: blume-20030912-various
2259 :     Description:
2260 :    
2261 :     - fixed bug in Real.toLargeInt
2262 :     - fixed bug in Posix.ProcEnv.times
2263 :     - changed inputLine functions to return an option
2264 :     - minor installer improvements / bugfixes
2265 :     - changed default @SMLalloc parameter for x86/celeron to 64k
2266 :    
2267 :     ----------------------------------------------------------------------
2268 :     Name: Matthias Blume (blume (at) tti - c (dot) org)
2269 : mblume 1356 Date: 2003/09/09 22:00:00 CDT
2270 :     Tag: Release_110_43
2271 :     Description:
2272 :    
2273 :     New working release 110.43. New bootfiles.
2274 :    
2275 :     ----------------------------------------------------------------------
2276 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org)
2277 : mblume 1354 Date: 2003/09/09 19:20:00 CDT
2278 :     Tag: blume-20030909-installer
2279 :     Description:
2280 :    
2281 :     Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
2282 :     Modified config/install.bat to take advantage of it. Also modified
2283 :     config/install.sh (and called it config/new-install.sh) to take advantage
2284 :     of it on Unix systems. (The SML code is (supposed to be) platform-
2285 :     independent.)
2286 :    
2287 :     The installer can now install everything under Win32
2288 :     as well as under *nix as long as it compiles.
2289 :    
2290 :     Other changes:
2291 :    
2292 :     - made CML compile again under Win32
2293 :     - made eXene compile under Win32 (by providing a fake structure UnixSock
2294 :     and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
2295 :     - fixed a bug in nowhere: it assumed that type OS.Process.status is the
2296 :     same as type int; under Win32 it isn't
2297 :     - fixed some slice-related problems in the win32-specific parts of CML
2298 :     - added a functor argument "sameVol" to os-path-fn.sml in the Basis
2299 :     (under Win32, the volume name is case-insensitive, and the
2300 :     OS.Path code compares volume names for equality)
2301 :    
2302 :     ----------------------------------------------------------------------
2303 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org)
2304 : mblume 1353 Date: 2003/09/08 11:55:00 CDT
2305 :     Tag: blume-20030908-fullpath
2306 :     Description:
2307 :    
2308 :     Made Win32 version of OS.FileSys.fullPath return current directory
2309 :     when given an empty string. This is what the spec says, and incidentally,
2310 :     CM depends on it. (CM otherwise goes into an infinite loop in certain
2311 :     cases when presented with the name of a non-existing .cm file.)
2312 :    
2313 :     ----------------------------------------------------------------------
2314 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org)
2315 : mblume 1350 Date: 2003/09/04 16:30:00 CDT
2316 : mblume 1352 Tag: blume-20030905-slices-etc
2317 : mblume 1350 Description:
2318 :    
2319 :     1. Changed interface to vectors and arrays in Basis to match
2320 :     (draft) Basis spec.
2321 :     2. Added signatures and implementations of slices according to
2322 :     Basis spec.
2323 :     3. Edited source code throughout the system to make it compile again
2324 :     under 1. and 2. (In some cases code had to be added to have it
2325 :     match the new signatures.)
2326 :     4. MLRISC should be backward-compatible: the copies of the originals
2327 :     of files that needed to change under 3. were retained, the .cm files
2328 :     check the compiler version number and use old versions when
2329 :     appropriate.
2330 :     5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
2331 :     dirstream -> string option (in accordance with Basis spec).
2332 :     6. When generating code that counts lines, ml-lex used function
2333 :     CharVector.foldli, taking advantage of its old interface.
2334 :     This has been replaced with the corresponding code from
2335 :     CharVectorSlice. (html-lex must be re-lexed!)
2336 :     7. BitArray in smlnj-lib/Util has been extended/modified to match the
2337 :     new MONO_ARRAY signature. (Do we need BitArraySlice?)
2338 :     8. Removed temporary additions (fromInternal, toInternal) from the
2339 :     (now obsolete) IntInf in smlnj-lib/Util.
2340 :     9. Cleaned up structure Byte.
2341 :     10. Added localOffset, scan, and fromString to Date (according to spec).
2342 :     Cleaned/corrected implementation of Date.
2343 :     (Still need to check for correctness; implement better canonicalizeDate.)
2344 :     11. Added "scan" to signature IEEE_REAL.
2345 :     12. Some improvements to IntInf [in particular: efficiency-hack for
2346 :     mod and rem when second operand is 2 (for parity checks).]
2347 :     13. Changed representation of type Time.time, using a single IntInf.int
2348 :     value counting microseconds. This considerably simplified the
2349 :     implementation of structure Time. We now support negative time
2350 :     values; scan and fromString handle signs.
2351 :     14. Functor PrimIO now takes two additional arguments (VectorSlice and
2352 :     ArraySlice).
2353 :    
2354 :     ----------------------------------------------------------------------
2355 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org)
2356 : mblume 1347 Date: 2003/08/28 17:00:00 CDT
2357 :     Tag: blume-20030828-intinf
2358 :     Description:
2359 :    
2360 :     This is a major update which comes with a version number bump
2361 :     (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
2362 :     and an implementation of IntInf in the Basis.
2363 :    
2364 :     There are a fairly large number of related changes and updates throughout
2365 :     the system:
2366 :    
2367 :     Basis:
2368 :     - Implemented IntInf.
2369 :     - Made LargeInt a projection of IntInf (by filtering through INTEGER).
2370 : mblume 1348 - Added some missing Real64 operations, most notably Real.toLargeInt.
2371 : mblume 1347 - Added FixedInt as a synonym for Int32.
2372 :    
2373 :     compiler:
2374 :     * Added support for a built-in intinf type.
2375 :     - literals
2376 :     - pattern matching
2377 :     - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
2378 :     - overloading on literals and operations
2379 :    
2380 :     This required adding a primitive type intinf, some additional
2381 :     primops, and implementations for several non-trivial intinf
2382 :     operations in Core. (The intinf type is completely abstract
2383 :     to the compiler; all operations get delegated back to the Core.)
2384 :    
2385 :     * Intinf equality is handled by polyequal. However, the compiler
2386 :     does not print its usual warning in this case (since polyequal
2387 :     is the right thing to do there).
2388 :    
2389 :     * Improved the organization of structure InlineT.
2390 :    
2391 :     * A word about conversion primops:
2392 :     If conversions involving intinf do not cancel out during
2393 : mblume 1348 CPS contract, then the compiler must insert calls to Core functions.
2394 : mblume 1347 Since all core access must be resolved already during the FLINT
2395 :     translate phase, it would be too late a the time of CPS contract
2396 : mblume 1348 to add new Core calls. For this reason, conversion primops
2397 : mblume 1347 for intinf carry two arguments: 1. the numeric argument that
2398 :     they are supposed to convert, and 2. the Core function that
2399 :     can help with this conversion if necessary. If CPS contract
2400 : mblume 1348 eliminates a primop, then the associated Core function becomes
2401 :     dead and goes away. Intinf conversion primops that do not get
2402 :     eliminated by CPS contract get rewritten into calls of their
2403 :     core functions by a separate, new phase.
2404 : mblume 1347
2405 :     interactive system:
2406 :     - Control.Print.intinfDepth controls max length of intinf constants
2407 :     being printed. (Analogous to Control.Print.stringDepth.)
2408 :     - Cleanup in printutil and pputil: got rid of unused stuff and
2409 :     duplicates; replaced some of the code with code that makes better
2410 :     use of library functionality.
2411 :    
2412 :     CM:
2413 :     Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
2414 :     are no longer silent.
2415 :    
2416 :     CKIT:
2417 :     Fixed mismatched uses of Int32 and LargeInt. I always decided
2418 :     in favor of LargeInt -- which is now the same as IntInf.
2419 :     CKIT-knowledgable people should check whether this is what's
2420 :     intended and otherwise change things back to using Int32 or
2421 :     FixedInt.
2422 :    
2423 :     Throughout the code:
2424 : mblume 1348 Started using IntInf.int literals and built-in operations
2425 : mblume 1347 (e.g., comparison with 0) where this seems appropriate.
2426 :    
2427 :    
2428 :     ----------------------------------------------------------------------
2429 : macqueen 1343 Name: Dave MacQueen (dbm@cs.uchicago.edu)
2430 :     Date: 2003/08/13 11:36:00 CDT
2431 :     Tag: dbm-20030813-mcz-merge1
2432 :     Description:
2433 :    
2434 :     Merging changes from the mcz-branch development branch into trunk.
2435 :     These changes involve replacement of the emulated old prettyprinter
2436 :     interface with direct use of the SML/NJ Lib PP library, and fixing
2437 :     of a couple of bugs (895, 1186) relating to error messages. A new
2438 :     prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
2439 :     has been added.
2440 :    
2441 :     ----------------------------------------------------------------------
2442 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org)
2443 : mblume 1342 Date: 2003/08/11 15:45:00 CDT
2444 :     Tag: blume-20030811-windows
2445 :     Description:
2446 :    
2447 :     Version number bumped to 110.42.9. NEW BOOTFILES!!!
2448 :    
2449 :     http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
2450 :    
2451 :     This patch restores SML/NJ's ability to run under win32. There are a
2452 :     number of changes, including fixes for several bugs that had gone
2453 :     unnoticed until now:
2454 :    
2455 :     - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
2456 :     trying to run the win32 version from within cygwin.)
2457 :    
2458 :     - There are a number of simple .bat scripts that substitute for their
2459 :     corresponding Unix shell-scripts. (See below.)
2460 :    
2461 :     - The internals of ml-build have been modified slightly. The main
2462 :     difference is that instead of calling ".link-sml" (or link-sml.bat)
2463 :     using OS.Process.system, the ML process delegates this task back
2464 :     to the script. Otherwise problems arise in mixed environments such
2465 :     as Cygwin where scripts look and work like Unix scripts, but
2466 :     where OS.Process.system cannot run them.
2467 :    
2468 :     - In CM, the srcpath pickler used native pathname syntax -- which
2469 :     is incorrect in the case of cross-compilation. The new pickle format
2470 :     is independent of platform-specific naming conventions.
2471 :    
2472 :     - Path configuration files (such as lib/pathconfig) can now choose
2473 :     between native and standard syntax. Placing a line of the form
2474 :    
2475 :     standard!
2476 :    
2477 :     into the file causes all subsequent paths to be interpreted using
2478 :     CM standard pathname syntax (= Unix conventions); a line
2479 :    
2480 :     native!
2481 :    
2482 :     switches back to native style. This was needed so that
2483 :     path config files can be written portably, see src/system/pathconfig.
2484 :    
2485 :     - Runtime system:
2486 :    
2487 :     - win32-filesys.c: get_file_time and set_file_time now
2488 :     access modification time, not creation time.
2489 :    
2490 :     - I/O code made aware of new array representation.
2491 :    
2492 :     - Bug fixes in X86.prim.masm.
2493 :    
2494 :     - src/system/makeml made aware of win32. (For use under cygwin
2495 :     and other Unix-environments for windows.)
2496 :    
2497 :     - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
2498 :     which caused BinIO.inputAll to fail consistently.
2499 :    
2500 :     .bat scripts:
2501 :    
2502 :     Windows .bat scripts assume that SMLNJ_HOME is defined.
2503 :    
2504 :     - sml.bat, ml-yacc.bat, ml-lex.bat: Driver scripts for standalone
2505 :     applications (sml, ml-yacc, ml-lex).
2506 :     - ml-build.bat: analogous to ml-build.
2507 :     - config\install.bat: Analogous to config/install.sh. This requires
2508 :     that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
2509 :     (nmake etc. must be on the path, and vcvars32 must have been run.)
2510 :     Moreover, sources for ml-lex and ml-yacc need to exist under src,
2511 :     and the bootfile hierarchy must have been unpacked under
2512 :     sml.boot.x86-win32.
2513 :     The script is very primitive and does a poor job at error checking.
2514 :     It only installs the base system, ml-lex, and ml-yacc. No other
2515 :     libraries are being installed (i.e., you get only those that
2516 :     are part of the compiler.)
2517 :     - link-sml.bat: analogous to .link-sml, but not currently used
2518 :    
2519 :     Unrelated bug fixes:
2520 :    
2521 :     - ml-nlffigen now exports structures ST_* corresponding to incomplete
2522 :     types.
2523 :     - Added getDevice to PP/src/pp-debug-fn.sml. (Would not compile
2524 :     otherwise.)
2525 :    
2526 :     ----------------------------------------------------------------------
2527 : macqueen 1337 Name: Dave MacQueen (macqueen@cs.uchicago.edu)
2528 :     Date: 2003/06/17
2529 :     Tag: macqueen-20030617-bug895
2530 :     Description:
2531 :    
2532 :     Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
2533 :     Tag will be used for new development branch (mcz-branch) for
2534 :     use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
2535 :    
2536 :     ----------------------------------------------------------------------
2537 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org)
2538 : mblume 1335 Date: 2003/05/27 16:55:00 CDT
2539 :     Tag: blume-20030527-polyeq
2540 :     Description:
2541 :    
2542 :     Tried to eliminated most cases of polymorphic equality.
2543 :    
2544 :     ----------------------------------------------------------------------
2545 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org)
2546 : mblume 1334 Date: 2003/05/21 17:45:00 CDT
2547 :     Tag: blume-20030517-complete
2548 :     Description:
2549 :    
2550 :     Two changes:
2551 :    
2552 :     1. Added a flag for controlling whether non-exhaustive bindings will
2553 :     be treated as errors (default is false).
2554 :     2. Cleaned up the *entire* source tree so that CMB.make goes through
2555 :     without a single non-exhaustive match- or bind warning.
2556 :    
2557 :     ----------------------------------------------------------------------
2558 : mblume 1368 Name: Matthias Blume (blume (at) tti - c (dot) org)
2559 : mblume 1332 Date: 2003/05/17 10:20:00 CDT
2560 :     Tag: blume-20030517-absyn
2561 :     Description:
2562 :    
2563 :     1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
2564 :    
2565 :     This mainly affects the quality of error messages. However, some
2566 :     of the code is now more straightforward than before. (Treatment of
2567 :     the above four constructs in translate.sml is much simpler than
2568 :     the "macro-expansion" that was going on before. Plus, the mach-
2569 :     compiler no longer gets invoked just to be able to compile an
2570 :     if-expression.)
2571 :    
2572 :     2. The ErrorMsg.Error exception is now caught and absorbed by the
2573 :     interactive loop.
2574 :    
2575 :     ----------------------------------------------------------------------
2576 : allenleung 1327 Name: Allen Leung
2577 :     Date: 2003/05/16 13:05:00 CDT
2578 :     Tag: leunga-20030516-cygwin-runtime
2579 :     Description:
2580 :    
2581 : blume 2155 Ported the runtime system to cygwin, which uses the unix
2582 : allenleung 1327 x86-unix bin files. Missing/buggy features:
2583 :    
2584 : blume 2155 o getnetbyname, getnetbyaddr: these functions seem to be missing in
2585 : allenleung 1327 the Cygwin library.
2586 :     o Ctrl-C handling may be flaky.
2587 :     o Windows system calls and Windows I/O are not supported.
2588 :    
2589 :     A new set of binfiles is located at:
2590 :    
2591 :     http://www.dorsai.org/~leunga/boot.x86-unix.tgz
2592 :    
2593 :     This is only needed for bootstrapping the cygwin version of smlnj.
2594 :     Other x86 versions can use the existing binfiles.
2595 :    
2596 :     ----------------------------------------------------------------------
2597 : mblume 1317 Name: Matthias Blume
2598 :     Date: 2003/04/08 15:42:00 CDT
2599 :     Tag: blume-20030408-listpair
2600 :     Description:
2601 :    
2602 :     1. Added a target 'mlrisc' to installer.
2603 :    
2604 :     2. Added missing elements to structure ListPair.
2605 :    
2606 :     ----------------------------------------------------------------------
2607 : leunga 1296 Name: Allen Leung
2608 :     Date: 2003/01/07 10:40:00 EST
2609 :     Tag: leunga-20030107-int-rem
2610 :     Description:
2611 :    
2612 :     Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
2613 :     The arguments to the SUBL instruction were swapped.
2614 :    
2615 :     ----------------------------------------------------------------------
2616 : blume 1229 Name: Matthias Blume
2617 : blume 1294 Date: 2002/12/12 16:25:00 EST
2618 :     Tag: blume-20021212-risc-ra
2619 :     Description:
2620 :    
2621 :     Fixed a serious bug in the rewrite code for FP spilling/reloading that
2622 :     sent the RA into an infinite loop when floating point registers get
2623 :     spilled. (Because of this bug, e.g., nucleic stopped compiling between
2624 :     110.37 and 110.38.)
2625 :     There was another set of potential problems related to the handling of
2626 :     MLRISC annotations (but those did not yet cause real problems, apparently).
2627 :    
2628 :     ----------------------------------------------------------------------
2629 :     Name: Matthias Blume
2630 : blume 1291 Date: 2002/12/06 22:40:00 EST
2631 :     Tag: blume-20021206-cm-fileid
2632 :     Description:
2633 :    
2634 :     Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
2635 :     This fixes the problem of CM getting confused by files that suddenly
2636 :     change their identity (e.g., by getting unlinked and recreated by some
2637 :     text editor such as vi). There might be a better/cheaper/cleaner way
2638 :     of doing this, but for now this will have to do.
2639 :    
2640 :     ----------------------------------------------------------------------
2641 :     Name: Matthias Blume
2642 : blume 1288 Date: 2002/10/28 09:50:00 EST
2643 : blume 2155 Tag: blume-20021028-typecheck
2644 : blume 1288 Description:
2645 :    
2646 :     Exported structure Typecheck from $smlnj/viscomp/core.cm.
2647 :    
2648 :     ----------------------------------------------------------------------
2649 :     Name: Matthias Blume
2650 : blume 1282 Date: 2002/10/17 09:10:00 EDT
2651 : blume 1279 Tag: Release_110_42
2652 :     Description:
2653 :    
2654 : blume 1282 In good old tradition, there has been a slight hiccup so that we have
2655 :     to patch 110.42 after the fact. The old release tag has been replaced
2656 :     (see below).
2657 :    
2658 :     The change solves a problem with two competing approaches the
2659 :     configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
2660 :     each other's way.
2661 :    
2662 :     This change only affects the runtime system code and the installer script.
2663 :     (No new bootfiles.)
2664 :    
2665 :     ----------------------------------------------------------------------
2666 :     Name: Matthias Blume
2667 :     Date: 2002/10/16 12:00:00 EDT
2668 :     Tag: Release_110_42_removed
2669 :     Description:
2670 :    
2671 : blume 1279 New working release. New bootfiles.
2672 :    
2673 :     ----------------------------------------------------------------------
2674 :     Name: Matthias Blume
2675 : blume 1273 Date: 2002/10/10 13:10:00 EDT
2676 :     Tag: blume-20021010-ppc-divs
2677 :     Description:
2678 :    
2679 :     The mltree operator DIVS must be implemented with an overflow check on
2680 :     the PPC because the hardware indicates divide-by-zero using "overflow" as
2681 :     well.
2682 :    
2683 :     ----------------------------------------------------------------------
2684 :     Name: Matthias Blume
2685 : blume 1261 Date: 2002/07/23 11:45:00
2686 :     Tag: blume-20020723-smlnj-home
2687 :     Description:
2688 :    
2689 :     Sml now senses the SMLNJ_HOME environment variable. If this is set,
2690 :     then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
2691 :     CM_PATHCONFIG is also set), the path configuration file is assumed
2692 :     to be in $SMLNJ_HOME/lib/pathconfig. This way one can easily move
2693 :     the entire tree to some other place and everything will "just work".
2694 :    
2695 :     (Companion commands such as ml-build and ml-makedepend also sense this
2696 :     variable.)
2697 :    
2698 :     ----------------------------------------------------------------------
2699 :     Name: Matthias Blume
2700 : blume 1259 Date: 2002/07/12 21:19:00 EDT
2701 :     Tag: blume-20020712-liveness
2702 :     Description:
2703 :    
2704 :     Exported two useful "step" functions from liveness module (MLRISC).
2705 :    
2706 :     ----------------------------------------------------------------------
2707 :     Name: Matthias Blume
2708 : blume 1253 Date: 2002/07/05 16:00 EDT
2709 :     Tag: Release_110_41
2710 :     Description:
2711 :    
2712 :     New working release. New bootfiles.
2713 :    
2714 :     ----------------------------------------------------------------------
2715 :     Name: Matthias Blume
2716 : blume 1250 Date: 2002/07/05 10:25:00 EDT
2717 :     Tag: blume-20020705-btimp
2718 :     Description:
2719 :    
2720 :     Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
2721 :     other clients can set up backtracing support.
2722 :    
2723 :     ----------------------------------------------------------------------
2724 :     Name: Matthias Blume
2725 : blume 1249 Date: 2002/06/25 17:23:00 EDT
2726 :     Tag: blume-20020625-fpmax
2727 :     Description:
2728 :    
2729 :     Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
2730 :     case. (The sense of the isNaN test was reversed -- which made min and
2731 :     max always return their first argument.)
2732 :    
2733 :     ----------------------------------------------------------------------
2734 :     Name: Matthias Blume
2735 : blume 1240 Date: 2002/06/11
2736 :     Tag: blume-20020611-unixpath
2737 :     Description:
2738 :    
2739 :     Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
2740 :    
2741 :     ----------------------------------------------------------------------
2742 :     Name: Matthias Blume
2743 : blume 1238 Date: 2002/06/10 16:35:00 EDT
2744 :     Tag: blume-20020610-ieeereal
2745 :     Description:
2746 :    
2747 :     I back-ported my implementation of IEEEReal.fromString from the
2748 :     idlbasis-devel branch so that we can test it.
2749 :    
2750 :     Another small change is that ppDec tries to give more information
2751 :     than just "<sig>" in the case of functors. However, this code is
2752 :     broken in some mysterious way if the functor's body's signature
2753 :     has not been declared by ascription but gets inferred from the
2754 :     implementation. This needs fixing...
2755 :    
2756 :     ----------------------------------------------------------------------
2757 :     Name: Matthias Blume
2758 : blume 1229 Date: 2002/05/31
2759 :     Tag: blume-20020531-btrace-mode
2760 :     Description:
2761 :    
2762 :     Resurrected SMLofNJ.Internals.BTrace.mode. (It accidentally fell by
2763 :     the wayside when I switched over to using Controls everywhere.)
2764 :    
2765 :     ----------------------------------------------------------------------
2766 : george 1222 Name: Lal George
2767 :     Date: 2002/05/23 12:21:40 EDT
2768 :     Tag: george-20020523-visual-labels
2769 :     Description:
2770 :    
2771 :     Labels are now displayed in the graphical output to make
2772 :     the fall-through and target blocks obvious.
2773 :    
2774 :     ----------------------------------------------------------------------
2775 : blume 1204 Name: Matthias Blume
2776 : blume 1220 Date: 2002/05/22 11:03:00 EDT
2777 :     Tag: blume-20020522-shrink
2778 :     Description:
2779 :    
2780 :     John tweaked yesterday's fix for 1131 to handle an out-of-memory
2781 :     situation that comes up when allocating huge arrays.
2782 :    
2783 :     ----------------------------------------------------------------------
2784 :     Name: Matthias Blume
2785 : blume 1215 Date: 2002/05/21 16:00:00 EDT
2786 :     Tag: Release_110_40
2787 :     Description:
2788 :    
2789 :     New working release (110.40). New bootfiles.
2790 :    
2791 :     [Also: John Reppy fixed GC bug 1131.]
2792 :    
2793 :     ----------------------------------------------------------------------
2794 :     Name: Matthias Blume
2795 : blume 1212 Date: 2002/05/21 12:35:00 EDT
2796 :     Tag: blume-20020521-cmdoc
2797 :     Description:
2798 :    
2799 :     CM documentation update.
2800 :    
2801 :     ----------------------------------------------------------------------
2802 :     Name: Matthias Blume
2803 : blume 1211 Date: 2002/05/21 10:55:00 EDT
2804 :     Tag: blume-20020521-misc
2805 :     Description:
2806 :    
2807 :     - John tweaked runtime to be silent on heap export (except when
2808 :     GC messages are on).
2809 :     - I added a few more things (cross-compiling versions of CMB) to
2810 :     config/preloads (as suggestions).
2811 :    
2812 :     ----------------------------------------------------------------------
2813 :     Name: Matthias Blume
2814 : blume 1208 Date: 2002/05/20 22:25:00 EDT
2815 :     Tag: blume-20020520-controls
2816 :     Description:
2817 :    
2818 :     - Added ControlUtil structure to control-lib.cm.
2819 :     - Use it throughout.
2820 :     - Used Controls facility to define MLRISC controls (as opposed to
2821 :     registering MLRISC control ref cells with Controls after the fact)
2822 :     - Fixed messed-up controls priorities.
2823 :    
2824 :     * Removed again all the stuff from config/preloads that one wouldn't
2825 :     be able to preload at the time the initial heap image is built.
2826 :     (Many libraries, e.g., CML, do not exist yet at this time. The
2827 :     only libraries that can be preloaded via config/preloads are those
2828 :     that come bundled with the bootfiles.)
2829 :    
2830 :     ----------------------------------------------------------------------
2831 :     Name: Matthias Blume
2832 : blume 1204 Date: 2002/05/20 10:59:00 EDT
2833 :     Tag: blume-20020520-preloads
2834 :     Description:
2835 :    
2836 :     Added a lot of commented-out suggestions for things to be included
2837 :     in config/preloads.
2838 :    
2839 :     ----------------------------------------------------------------------
2840 : leunga 1203 Name: Allen Leung
2841 :     Date: 2002/05/18 14:20:00 EDT
2842 :     Tag: leunga-20020518-mdl
2843 :     Description:
2844 :    
2845 :     o Made the mdl tool stuff compile and run again.
2846 :     o I've disabled all the stuff that depends on RTL specifications; they
2847 :     are all badly broken anyway.
2848 :    
2849 :     ----------------------------------------------------------------------
2850 : blume 1201 Name: Matthias Blume
2851 :     Date: 2002/05/17 16:49:00 EDT
2852 :     Tag: blume-20020517-controls
2853 :     Description:
2854 :    
2855 :     0. John Reppy made several modifications to the SML/NJ library.
2856 :     In particular, there is a shiny new controls-lib.cm.
2857 :    
2858 :     1. Pushed new controls interface through compiler so that everything
2859 :     compiles again.
2860 :    
2861 :     2. Added FormatComb and FORMAT_COMB to the CML version of the
2862 :     SML/NJ library (so that CML compiles again).
2863 :    
2864 :     3. Modified init scripts because XXX_DEFAULT environment variables
2865 :     are no longer with us. (Boot-time initialization is now done
2866 :     using the same environment variables that are also used for
2867 :     startup-time initialization of controls.)
2868 :    
2869 :     ----------------------------------------------------------------------
2870 : george 1192 Name: Lal George
2871 :     Date: 2002/05/15 09:20:10 EDT
2872 :     Tag: george-20020515-pseudo-op-decls
2873 :     Description:
2874 :    
2875 :     All pseudo-ops emitted before the first segment declaration
2876 :     such as TEXT, DATA, and BSS directives are assumed to be global
2877 :     declarations and are emitted first in the assembly file. This is
2878 :     useful in a number of situations where one has pseudo-ops that are not
2879 :     specific to any segment, and also works around the constraint that one
2880 :     cannot have client pseudo-ops in the TEXT segment.
2881 :    
2882 :     Because no segment is associated with these declarations it is
2883 :     an error to allocate any space or objects before the first segment
2884 :     directive and an exception will be raised. However, we cannot make
2885 :     this check for client pseudo-ops.
2886 :    
2887 :     These top level declarations are a field in the CFG graph_info.
2888 :     In theory you can continue to add to this field after the CFG has been
2889 :     built -- provided you know what you are doing;-)
2890 :    
2891 :     ----------------------------------------------------------------------
2892 : blume 1175 Name: Matthias Blume
2893 : blume 1190 Date: 2002/05/13 16:40:00 EDT
2894 :     Tag: blume-20020513-pp-etc
2895 :     Description:
2896 :    
2897 :     A few minor bugfixes:
2898 :    
2899 :     - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
2900 :     (Bogus printouts for redefined bindings still occur. Compiler
2901 :     bug should no longer occur now. We need to redo the prettyprinter
2902 :     from scratch.)
2903 :    
2904 :     - CM pathname printer now also adds escape sequences for ( and )
2905 :    
2906 :     - commend and docu fixes for ml-nlffi
2907 :    
2908 :     ----------------------------------------------------------------------
2909 :     Name: Matthias Blume
2910 : blume 1189 Date: 2002/05/10 16:40:00 EDT
2911 :     Tag: blume-20020510-erg-textio
2912 :     Description:
2913 :    
2914 :     Applied the following bugfix provided by Emden Gansner:
2915 :    
2916 :     Output is corrupted when outputSubstr is used rather than output.
2917 :    
2918 : blume 2155 The problem occurs when a substring
2919 : blume 1189
2920 :     ss = (s, dataStart, dataLen)
2921 :    
2922 : blume 2155 where dataStart > 0, fills a stream buffer with avail bytes left.
2923 :     avail bytes of s, starting at index dataStart, are copied into the
2924 : blume 1189 buffer, the buffer is flushed, and then the remaining dataLen-avail
2925 : blume 2155 bytes of ss are copied into the beginning of the buffer. Instead of
2926 :     starting this copy at index dataStart+avail in s, the current code
2927 : blume 1189 starts the copy at index avail.
2928 :    
2929 :     Fix:
2930 :     In text-io-fn.sml, change line 695 from
2931 :     val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2932 : blume 2155 to
2933 : blume 1189 val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2934 :    
2935 :     ----------------------------------------------------------------------
2936 :     Name: Matthias Blume
2937 : blume 1186 Date: 2002/04/12 13:55:00 EDT
2938 :     Tag: blume-20020412-assyntax
2939 :     Description:
2940 :    
2941 :     1. Grabbed newer assyntax.h from the XFree86 project.
2942 :     2. Fiddled with how to compile X86.prim.asm without warnings.
2943 :     3. (Very) Minor cleanup in CM.
2944 :    
2945 :     ----------------------------------------------------------------------
2946 :     Name: Matthias Blume
2947 : blume 1185 Date: 2002/04/01 (no joke!) 17:07:00 EST
2948 :     Tag: blume-20020401-x86div
2949 :     Description:
2950 :    
2951 :     Added full support for div/mod/rem/quot on the x86, using the machine
2952 :     instruction's two results (without clumsily recomputing the remainder)
2953 :     directly where appropriate.
2954 :    
2955 :     Some more extensive power-of-two support was added to the x86 instruction
2956 :     selector (avoiding expensive divs, mods, and muls where they can be
2957 :     replaced with cheaper shifts and masks). However, this sort of thing
2958 :     ought to be done earlier, e.g., within the CPS optimizer so that
2959 :     all architectures benefit from it.
2960 :    
2961 :     The compiler compiles to a fixed point, but changes might be somewhat
2962 :     fragile nevertheless. Please, report any strange things that you might
2963 :     see wrt. div/mod/quot/rem...
2964 :    
2965 :     ----------------------------------------------------------------------
2966 :     Name: Matthias Blume
2967 : blume 1184 Date: 2002/03/29 17:22:00
2968 :     Tag: blume-20020329-div
2969 :     Description:
2970 :    
2971 :     Fixed my broken div/mod logic. Unfortunately, this means that the
2972 :     inline code for div/mod now has one more comparison than before.
2973 :     Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2974 :     The problem was with quotient = 0, because that alone does not tell
2975 :     us which way the rounding went. One then has to look at whether
2976 :     remainder and divisor have the same sign... :(
2977 :    
2978 :     Anyway, I replaced the bootfiles with fresh ones...
2979 :    
2980 :     ----------------------------------------------------------------------
2981 :     Name: Matthias Blume
2982 : blume 1183 Date: 2002/03/29 14:10:00 EST
2983 :     Tag: blume-20020329-inlprims
2984 :     Description:
2985 :    
2986 :     NEW BOOTFILES!!! Version number bumped to 110.39.3.
2987 :    
2988 :     Primops have changed. This means that the bin/boot-file formats have
2989 :     changed as well.
2990 :    
2991 :     To make sure that there is no confusion, I made a new version.
2992 :    
2993 :    
2994 :     CHANGES:
2995 :    
2996 :     * removed REMT from mltree (remainder should never overflow).
2997 :    
2998 :     * added primops to deal with divisions of all flavors to the frontend
2999 :    
3000 :     * handled these primops all the way through so they map to their respective
3001 :     MLRISC support
3002 :    
3003 :     * used these primops in the implementation of Int, Int32, Word, Word32
3004 :    
3005 :     * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
3006 :    
3007 :     * parameterized INLMIN, INLMAX, and INLABS by a numkind
3008 :    
3009 :     * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
3010 :     floating point
3011 :    
3012 :     * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
3013 :     and Real (but Real.abs maps to a separate floating-point-only primop)
3014 :    
3015 :    
3016 :     TODO items:
3017 :    
3018 :     * Hacked Alpha32 instruction selection, disabling the selection of REMx
3019 :     instructions because the machine instruction encoder cannot handle
3020 :     them. (Hppa, PPC, and Sparc instruction selection did not handle
3021 :     REM in the first place, and REM is supported by the x86 machine coder.)
3022 :    
3023 :     * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
3024 :     selection phase. (The two can be streamlined because the hardware
3025 :     delivers both quotient and remainder at the same time anyway.)
3026 :    
3027 :     * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
3028 :     (Currently the behavior is inconsistent both across architectures and
3029 :     wrt. the draft Basis spec.)
3030 :    
3031 :     * Word8 should eventually be handled natively, too.
3032 :    
3033 :     * There seems to be one serious bug in mltree-gen.sml. It appears, though,
3034 :     as if there currently is no execution path that could trigger it in
3035 :     SML/NJ. (The assumptions underlying functions arith and promotable do not
3036 :     hold for things like multiplication and division.)
3037 :    
3038 :     ----------------------------------------------------------------------
3039 :     Name: Matthias Blume
3040 : blume 1181 Date: 2002/03/27 16:27:00 EST
3041 :     Tag: blume-20020327-mlrisc-divisions
3042 :     Description:
3043 :    
3044 :     Added support for all four division operations (ML's div, mod, quot,
3045 :     and rem) to MLRISC. In the course of doing so, I also rationalized
3046 :     the naming (no more annoying switch-around of DIV and QUOT), by
3047 :     parameterizing the operation by div_rounding_mode (which can be either
3048 :     DIV_TO_ZERO or DIV_TO_NEGINF).
3049 :    
3050 :     The generic MLTreeGen functor takes care of compiling all four
3051 :     operations down to only round-to-zero div.
3052 :    
3053 :     Missing pieces:
3054 :    
3055 :     * Doing something smarter than relying on MLTreeGen on architectures
3056 :     like, e.g., the x86 where hardware division delivers both quotient and
3057 :     remainder at the same time. With this, the implementation of the
3058 :     round-to-neginf operations could be further streamlined.
3059 :    
3060 :     * Remove inlining support for div/mod/rem from the frontend and replace it
3061 :     with primops that get carried through to the backend. Do this for all
3062 :     int and word types.
3063 :    
3064 :     ----------------------------------------------------------------------
3065 :     Name: Matthias Blume
3066 : blume 1180 Date: 2002/03/25 17:25:00 EST
3067 :     Tag: blume-20020325-divmod
3068 :     Description:
3069 :    
3070 :     I improved (hopefully without breaking them) the implementation of Int.div,
3071 :     Int.mod, and Int.rem. For this, the code in translate.sml now takes
3072 :     advantage of the following observations:
3073 :    
3074 :     Let q = x quot y r = x rem y
3075 :     d = x div y m = x mod y
3076 :    
3077 :     where "quot" is the round-to-zero version of integer division that
3078 :     hardware usually provides. Then we have:
3079 :    
3080 :     r = x - q * y where neither the * nor the - will overflow
3081 :     d = if q >= 0 orelse x = q * y then q else q - 1
3082 :     where neither the * nor the - will overflow
3083 :     m = if q >= 0 orelse r = 0 then r else r + y
3084 :     where the + will not overflow
3085 :    
3086 :     This results in substantial simplification of the generated code.
3087 :     The following table shows the number of CFG nodes and edges generated
3088 :     for
3089 :     fun f (x, y) = x OPER y
3090 :     (* with OPER \in div, mod, quot, rem *)
3091 :    
3092 :    
3093 :     OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
3094 :     --------------------------------------------------------
3095 :     div | 24 | 39 | 12 | 16
3096 :     mod | 41 | 71 | 12 | 16
3097 :     quot | 8 | 10 | 8 | 10
3098 :     rem | 10 | 14 | 8 | 10
3099 :    
3100 :    
3101 :     ----------------------------------------------------------------------
3102 :     Name: Matthias Blume
3103 : blume 1179 Date: 2002/03/25 22:06:00 EST
3104 :     Tag: blume-20020325-cprotobug
3105 :     Description:
3106 :    
3107 :     Fixed a bug in cproto (c prototype decoder).
3108 :    
3109 :     ----------------------------------------------------------------------
3110 :     Name: Matthias Blume
3111 : blume 1178 Date: 2002/03/25 16:00:00 EST
3112 :     Tag: blume-20020325-raw-primops
3113 :     Description:
3114 :    
3115 :     I did some cleanup to Allen's new primop code and
3116 :     replaced yesterday's bootfiles with new ones.
3117 :     (But they are stored in the same place.)
3118 :    
3119 :     ----------------------------------------------------------------------
3120 :     Name: Matthias Blume
3121 : blume 1175 Date: 2002/03/24 22:40:00 EST
3122 :     Tag: blume-20020324-bootfiles
3123 :     Description:
3124 :    
3125 :     Made the bootfiles that Allen asked for.
3126 :    
3127 :     ----------------------------------------------------------------------
3128 : leunga 1174 Name: Allen Leung
3129 :     Date: 2002/03/23 15:50:00 EST
3130 :     Tag: leunga-20020323-flint-cps-rcc-primops
3131 :     Description:
3132 :    
3133 :     1. Changes to FLINT primops:
3134 :    
3135 :     (* make a call to a C-function;
3136 :     * The primop carries C function prototype information and specifies
3137 :     * which of its (ML-) arguments are floating point. C prototype
3138 :     * information is for use by the backend, ML information is for
3139 :     * use by the CPS converter. *)
3140 :     | RAW_CCALL of { c_proto: CTypes.c_proto,
3141 :     ml_args: ccall_type list,
3142 :     ml_res_opt: ccall_type option,
3143 :     reentrant : bool
3144 :     } option
3145 :     (* Allocate uninitialized storage on the heap.
3146 :     * The record is meant to hold short-lived C objects, i.e., they
3147 : blume 2155 * are not ML pointers. With the tag, the representation is
3148 : leunga 1174 * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
3149 :     *)
3150 : blume 2155 | RAW_RECORD of {tag:bool,sz:int}
3151 : leunga 1174 and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
3152 :    
3153 :     2. These CPS primops are now overloaded:
3154 :    
3155 :     rawload of {kind:numkind}
3156 :     rawstore of {kind:numkind}
3157 :    
3158 :     The one argument form is:
3159 :    
3160 :     rawload {kind} address
3161 :    
3162 :     The two argument form is:
3163 : blume 2155
3164 : leunga 1174 rawload {kind} [ml object, byte-offset]
3165 :    
3166 :     3. RAW_CCALL/RCC now takes two extra arguments:
3167 :    
3168 :     a. The first is whether the C call is reentrant, i.e., whether
3169 :     ML state should be saved and restored.
3170 :     b. The second argument is a string argument specifying the name of
3171 :     library and the C function.
3172 :    
3173 :     These things are currently not handled in the code generator, yet.
3174 :    
3175 :     4. In CProto,
3176 :    
3177 : blume 2155 An encoding type of "bool" means "ml object" and is mapped into
3178 :     C prototype of PTR. Note that "bool" is different than "string",
3179 : leunga 1174 even though "string" is also mapped into PTR, because "bool"
3180 :     is assigned an CPS type of BOGt, while "string" is assigned INT32t.
3181 :    
3182 :     5. Pickler/unpicker
3183 :    
3184 :     Changed to handle RAW_RECORD and newest RAW_CCALL
3185 :    
3186 :     6. MLRiscGen,
3187 :    
3188 :     1. Changed to handle the new rawload/rawstore/rawrecord operators.
3189 :     2. Code for handling C Calls has been moved to a new module CPSCCalls,
3190 :     in the file CodeGen/cpscompile/cps-c-calls.sml
3191 :    
3192 :     7. Added the conditional move operator
3193 :    
3194 :     condmove of branch
3195 :    
3196 : blume 2155 to cps. Generation of this is still buggy so it is currently
3197 : leunga 1174 disabled.
3198 : blume 2155
3199 : leunga 1174 ----------------------------------------------------------------------
3200 : george 1168 Name: Lal George
3201 : blume 2155 Date: 2002/03/22 14:18:25 EST
3202 : george 1171 Tag: george-20020322-cps-branch-prob
3203 : george 1168 Description:
3204 :    
3205 : blume 2155 Implemented the Ball-Larus branch prediction-heuristics, and
3206 : george 1168 incorporated graphical viewers for control flow graphs.
3207 :    
3208 : george 1171 Ball-Larus Heuristics:
3209 : george 1168 ---------------------
3210 :     See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
3211 :    
3212 :     By design it uses the Dempster-Shafer theory for combining
3213 :     probabilities. For example, in the function:
3214 :    
3215 :     fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
3216 :    
3217 :     the ball-larus heuristics predicts that the n=0 is unlikely
3218 :     (OH-heuristic), and the 'then' branch is unlikely because of the
3219 :     RH-heuristic -- giving the 'then' branch an even lower combined
3220 : george 1171 probability using the Dempster-Shafer theory.
3221 : george 1168
3222 :     Finally, John Reppy's loop analysis in MLRISC, further lowers the
3223 :     probability of the 'then' branch because of the loop in the else
3224 :     branch.
3225 :    
3226 :    
3227 :     Graphical Viewing:
3228 : blume 2155 ------------------
3229 : george 1168 I merely plugged in Allen's graphical viewers into the compiler. The
3230 :     additional code is not much. At the top level, saying:
3231 :    
3232 :     Control.MLRISC.getFlag "cfg-graphical-view" := true;
3233 :    
3234 :     will display the graphical view of the control flow graph just before
3235 :     back-patching. daVinci must be in your path for this to work. If
3236 :     daVinci is not available, then the default viewer can be changed
3237 :     using:
3238 :    
3239 :     Control.MLRISC.getString "viewer"
3240 :    
3241 :     which can be set to "dot" or "vcg" for the corresponding viewers. Of
3242 :     course, these viewers must be in your path.
3243 :    
3244 :     The above will display the compilation unit at the level of clusters,
3245 :     many of which are small, boring, and un-interesting. Also setting:
3246 :    
3247 :     Control.MLRISC.getInt "cfg-graphical-view_size"
3248 :    
3249 :     will display clusters that are larger than the value set by the above.
3250 :    
3251 : blume 2155
3252 : george 1168 ----------------------------------------------------------------------
3253 : blume 1159 Name: Matthias Blume
3254 :     Date: 2002/03/21 22:20:00 EST
3255 :     Tag: blume-20020321-kmp-bugfix
3256 :     Description:
3257 :    
3258 :     Changed the interface to the KMP routine in PreString and fixed
3259 :     a minor bug in one place where it was used.
3260 :    
3261 :     ----------------------------------------------------------------------
3262 : leunga 1156 Name: Allen Leung
3263 : leunga 1158 Date: 2002/03/21 20:30:00 EST
3264 :     Tag: leunga-20020321-cfg
3265 :     Description:
3266 :    
3267 :     Fixed a potential problem in cfg edge splitting.
3268 :    
3269 :     ----------------------------------------------------------------------
3270 :     Name: Allen Leung
3271 : leunga 1156 Date: 2002/03/21 17:15:00 EST
3272 :     Tag: leunga-20020321-x86-fp-cfg
3273 :     Description:
3274 :    
3275 : blume 2155 1. Recoded the buggy parts of x86-fp.
3276 : leunga 1156
3277 :     a. All the block reordering code has been removed.
3278 :     We now depend on the block placement phases to do this work.
3279 :    
3280 :     b. Critical edge splitting code has been simplified and moved into the
3281 :     CFG modules, as where they belong.
3282 :    
3283 :     Both of these were quite buggy and complex. The code is now much, much
3284 :     simpler.
3285 :    
3286 :     2. X86 backend.
3287 :    
3288 :     a. Added instructions for 64-bit support. Instruction selection for
3289 :     64-bit has not been committed, however, since that
3290 : blume 2155 requires changes to MLTREE which haven't been approved by
3291 : leunga 1156 Lal and John.
3292 :    
3293 : blume 2155 b. Added support for FUCOMI and FUCOMIP when generating code for
3294 : leunga 1156 PentiumPro and above. We only generate these instructions in
3295 :     the fast-fp mode.
3296 :    
3297 : blume 2155 c. Added cases for JP and JNP in X86FreqProps.
3298 :    
3299 : leunga 1156 3. CFG
3300 :    
3301 :     CFG now has a bunch of methods for edge splitting and merging.
3302 :    
3303 :     4. Machine description.
3304 :    
3305 :     John's simplification of MLTREE_BASIS.fcond broke a few machine
3306 :     description things:
3307 :    
3308 :     rtl-build.{sig,sml} and hppa.mdl fixed.
3309 :    
3310 : blume 2155 NOTE: the machine description stuff in the repository is still broken.
3311 :     Again, I can't put my fixes in because that involves
3312 :     changes to MLTREE.
3313 : leunga 1156
3314 :     ----------------------------------------------------------------------
3315 : blume 1145 Name: Matthias Blume
3316 : blume 1155 Date: 2002/03/20 15:55:00 EST
3317 :     Tag: blume-20020320-kmp
3318 :     Description:
3319 :    
3320 :     Implemented Knuth-Morris-Pratt string matching in PreString and used
3321 :     it for String.isSubstring, Substring.isSubstring, and
3322 :     Substring.position.
3323 :    
3324 :     (Might need some stress-testing. Simple examples worked fine.)
3325 :    
3326 :     ----------------------------------------------------------------------
3327 :     Name: Matthias Blume
3328 : blume 1152 Date: 2002/03/19 16:37:00 EST
3329 :     Tag: blume-20020319-witnesses
3330 :     Description:
3331 :    
3332 :     Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
3333 :    
3334 :     This implements a generic mechanism for changing constness qualifiers
3335 :     anywhere within big C types without resorting to outright "casts".
3336 :     (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
3337 :     modify the constness at the outermost level.)
3338 :     The implementation of "convert" is based on the idea of "witness"
3339 :     values -- values that are not used by the operation but whose types
3340 :     "testify" to their applicability. On the implementation side, "convert"
3341 :     is simply a projection (returning its second curried argument). With
3342 :     cross-module inlining, it should not result in any machine code being
3343 :     generated.
3344 :    
3345 :     ----------------------------------------------------------------------
3346 :     Name: Matthias Blume
3347 : blume 1148 Date: 2002/03/15 16:40:00 EST
3348 :     Tag: blume-20020315-basis
3349 :     Description:
3350 :    
3351 :     Provided (preliminary?) implementations for
3352 :    
3353 :     {String,Substring}.{concatWith,isSuffix,isSubstring}
3354 :    
3355 :     and
3356 :    
3357 :     Substring.full
3358 :    
3359 :     Those are in the Basis spec but they were missing in SML/NJ.
3360 :    
3361 :     ----------------------------------------------------------------------
3362 :     Name: Matthias Blume
3363 : blume 1145 Date: 2002/03/14 21:30:00 EST
3364 :     Tag: blume-20020314-controls
3365 :     Description:
3366 :    
3367 :     Controls:
3368 :     ---------
3369 :    
3370 :     1. Factored out the recently-added Controls : CONTROLS stuff and put
3371 :     it into its own library $/controls-lib.cm. The source tree for
3372 :     this is under src/smlnj-lib/Controls.
3373 :    
3374 :     2. Changed the names of types and functions in this interface, so they
3375 :     make a bit more "sense":
3376 :    
3377 :     module -> registry
3378 :     'a registry -> 'a group
3379 :    
3380 :     3. The interface now deals in ref cells only. The getter/setter interface
3381 :     is (mostly) gone.
3382 :    
3383 :     4. Added a function that lets one register an already-existing ref cell.
3384 :    
3385 :     5. Made the corresponding modifications to the rest of the code so that
3386 :     everything compiles again.
3387 :    
3388 :     6. Changed the implementation of Controls.MLRISC back to something closer
3389 :     to the original. In particular, this module (and therefore MLRISC)
3390 :     does not depend on Controls. There now is some link-time code in
3391 :     int-sys.sml that registers the MLRISC controls with the Controls
3392 :     module.
3393 :    
3394 :     CM:
3395 :     ---
3396 :    
3397 :     * One can now specify the lambda-split aggressiveness in init.cmi.
3398 :    
3399 :     ----------------------------------------------------------------------
3400 : leunga 1142 Name: Allen Leung
3401 :     Date: 2002/03/13 17:30:00 EST
3402 :     Tag: leunga-20020313-x86-fp-unary
3403 :     Description:
3404 :    
3405 :     Bug fix for:
3406 :    
3407 :     > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
3408 :     > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
3409 :     > - fun f(x,(y,z)) = Real.~ y;
3410 :     > [autoloading]
3411 :     > [autoloading done]
3412 :     > fchsl (%eax), 184(%esp)
3413 :     > Error: MLRisc bug: X86MCEmitter.emitInstr
3414 : blume 2155 >
3415 : leunga 1142 > uncaught exception Error
3416 :     > raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
3417 :    
3418 :     The problem was that the code generator did not generate any fp registers
3419 :     in this case, and the ra didn't know that it needed to run the X86FP phase to
3420 :     translate the pseudo fp instruction. This only happened with unary fp
3421 :     operators in certain situations.
3422 :    
3423 :     ----------------------------------------------------------------------
3424 : blume 1137 Name: Matthias Blume
3425 : blume 1140 Date: 2002/03/13 14:00:00 EST
3426 :     Tag: blume-20020313-overload-etc
3427 :     Description:
3428 :    
3429 :     1. Added _overload as a synonym for overload for backward compatibility.
3430 :     (Control.overloadKW must be true for either version to be accepted.)
3431 :    
3432 :     2. Fixed bug in install script that caused more things to be installed
3433 :     than what was requested in config/targets.
3434 :    
3435 :     3. Made CM aware of the (_)overload construct so that autoloading
3436 :     works.
3437 :    
3438 :     ----------------------------------------------------------------------
3439 :     Name: Matthias Blume
3440 : blume 1138 Date: 2002/03/12 22:03:00 EST
3441 :     Tag: blume-20020312-url
3442 :     Description:
3443 :    
3444 :     Forgot to update BOOT and srcarchiveurl.
3445 :    
3446 :     ----------------------------------------------------------------------
3447 :     Name: Matthias Blume
3448 : blume 1137 Date: 2002/03/12 17:30:00 EST
3449 :     Tag: blume-20020312-version110392
3450 :     Description:
3451 :    
3452 :     Yet another version number bump (because of small changes to the
3453 :     binfile format). Version number is now 110.39.2. NEW BOOTFILES!
3454 :    
3455 :     Changes:
3456 :    
3457 :     The new pid generation scheme described a few weeks ago was overly
3458 :     complicated. I implemented a new mechanism that is simpler and
3459 :     provides a bit more "stability": Once CM has seen a compilation
3460 :     unit, it keeps its identity constant (as long as you do not delete
3461 :     those crucial CM/GUID/* files). This means that when you change
3462 : blume 1145 an interface, compile, then go back to the old interface, and
3463 : blume 1137 compile again, you arrive at the original pid.
3464 :    
3465 :     There now also is a mechanism that instructs CM to use the plain
3466 :     environment hash as a module's pid (effectively making its GUID
3467 :     the empty string). For this, "noguid" must be specified as an
3468 :     option to the .sml file in question within its .cm file.
3469 :     This is most useful for code that is being generated by tools such
3470 :     as ml-nlffigen (because during development programmers tend to
3471 :     erase the tool's entire output directory tree including CM's cached
3472 :     GUIDs). "noguid" is somewhat dangerous (since it can be used to locally
3473 :     revert to the old, broken behavior of SML/NJ, but in specific cases
3474 :     where there is no danger of interface confusion, its use is ok
3475 :     (I think).
3476 :    
3477 :     ml-nlffigen by default generates "noguid" annotations. They can be
3478 :     turned off by specifying -guid in its command line.
3479 :    
3480 :     ----------------------------------------------------------------------
3481 : george 1136 Name: Lal George
3482 :     Date: 2002/03/12 12 14:42:36 EST
3483 :     Tag: george-20020312-frequency-computation
3484 :     Description:
3485 : blume 2155
3486 : george 1136 Integrated jump chaining and static block frequency into the
3487 :     compiler. More details and numbers later.
3488 : leunga 1025
3489 : george 1003 ----------------------------------------------------------------------
3490 : george 1133 Name: Lal George
3491 :     Date: 2002/03/11 11 22:38:53 EST
3492 :     Tag: george-20020311-jump-chain-elim
3493 :     Description:
3494 :    
3495 :     Tested the jump chain elimination on all architectures (except the
3496 :     hppa). This is on by default right now and is profitable for the
3497 :     alpha and x86, however, it may not be profitable for the sparc and ppc
3498 :     when compiling the compiler.
3499 :    
3500 : blume 2155 The gc test will typically jump to a label at the end of the cluster,
3501 :     where there is another jump to an external cluster containing the actual
3502 : george 1133 code to invoke gc. This is to allow factoring of common gc invocation
3503 :     sequences. That is to say, we generate:
3504 :    
3505 :     f:
3506 :     testgc
3507 :     ja L1 % jump if above to L1
3508 :    
3509 :     L1:
3510 :     jmp L2
3511 :    
3512 :    
3513 :     After jump chain elimination the 'ja L1' instructions is converted to
3514 :     'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
3515 :     up being implemented in their long form (if L2 is far away) using:
3516 :    
3517 :     jbe L3 % jump if below or equal to L3
3518 :     jmp L2
3519 :     L3:
3520 :     ...
3521 :    
3522 :    
3523 :     For large compilation units L2 may be far away.
3524 :    
3525 :    
3526 :     ----------------------------------------------------------------------
3527 : blume 1132 Name: Matthias Blume
3528 :     Date: 2002/03/11 13:30:00 EST
3529 :     Tag: blume-20020311-mltreeeval
3530 :     Description:
3531 :    
3532 :     A functor parameter was missing.
3533 :    
3534 :     ----------------------------------------------------------------------
3535 : leunga 1130 Name: Allen Leung
3536 : leunga 1131 Date: 2002/03/11 10:30:00 EST
3537 : leunga 1142 Tag: leunga-20020311-runtime-string0
3538 : leunga 1131 Description:
3539 :    
3540 : blume 2155 The representation of the empty string now points to a
3541 :     legal null terminated C string instead of unit. It is now possible
3542 :     to convert an ML string into C string with InlineT.CharVector.getData.
3543 : leunga 1131 This compiles into one single machine instruction.
3544 :    
3545 :     ----------------------------------------------------------------------
3546 :     Name: Allen Leung
3547 : leunga 1130 Date: 2002/03/10 23:55:00 EST
3548 :     Tag: leunga-20020310-x86-call
3549 :     Description:
3550 :    
3551 :     Added machine generation for CALL instruction (relative displacement mode)
3552 :    
3553 :     ----------------------------------------------------------------------
3554 : blume 1128 Name: Matthias Blume
3555 :     Date: 2002/03/08 16:05:00
3556 :     Tag: blume-20020308-entrypoints
3557 :     Description:
3558 :    
3559 :     Version number bumped to 110.39.1. NEW BOOTFILES!
3560 :    
3561 :     Entrypoints: non-zero offset into a code object where execution should begin.
3562 :    
3563 :     - Added the notion of an entrypoint to CodeObj.
3564 :     - Added reading/writing of entrypoint info to Binfile.
3565 :     - Made runtime system bootloader aware of entrypoints.
3566 :     - Use the address of the label of the first function given to mlriscGen
3567 :     as the entrypoint. This address is currently always 0, but it will
3568 :     not be 0 once we turn on block placement.
3569 :     - Removed the linkage cluster code (which was The Other Way(tm) of dealing
3570 :     with entry points) from mlriscGen.
3571 :    
3572 :     ----------------------------------------------------------------------
3573 : leunga 1127 Name: Allen Leung
3574 :     Date: 2002/03/07 20:45:00 EST
3575 :     Tag: leunga-20020307-x86-cmov
3576 :     Description:
3577 :    
3578 :     Bug fixes for CMOVcc on x86.
3579 :    
3580 :     1. Added machine code generation for CMOVcc
3581 :     2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
3582 :     3. CMOVcc cannot have an immediate operand as argument.
3583 :    
3584 :     ----------------------------------------------------------------------
3585 : blume 1126 Name: Matthias Blume
3586 :     Date: 2002/03/07 16:15:00 EST
3587 :     Tag: blume-20020307-controls
3588 :     Description:
3589 :    
3590 :     This is a very large but mostly boring patch which makes (almost)
3591 :     every tuneable compiler knob (i.e., pretty much everything under
3592 :     Control.* plus a few other things) configurable via both the command
3593 :     line and environment variables in the style CM did its configuration
3594 :     until now.
3595 :    
3596 :     Try starting sml with '-h' (or, if you are brave, '-H')
3597 :    
3598 :     To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
3599 :     implements the underlying generic mechanism.
3600 :    
3601 :     The interface to some of the existing such facilities has changed somewhat.
3602 :     For example, the MLRiscControl module now provides mkFoo instead of getFoo.
3603 :     (The getFoo interface is still there for backward-compatibility, but its
3604 :     use is deprecated.)
3605 :    
3606 :     The ml-build script passes -Cxxx=yyy command-line arguments through so
3607 :     that one can now twiddle the compiler settings when using this "batch"
3608 :     compiler.
3609 :    
3610 :     TODO items:
3611 :    
3612 :     We should go through and throw out all controls that are no longer
3613 :     connected to anything. Moreover, we should go through and provide
3614 :     meaningful (and correct!) documentation strings for those controls
3615 :     that still are connected.
3616 :    
3617 :     Currently, multiple calls to Controls.new are accepted (only the first
3618 :     has any effect). Eventually we should make sure that every control
3619 :     is being made (via Controls.new) exactly once. Future access can then
3620 :     be done using Controls.acc.
3621 :    
3622 :     Finally, it would probably be a good idea to use the getter-setter
3623 :     interface to controls rather than ref cells. For the time being, both
3624 :     styles are provided by the Controls module, but getter-setter pairs are
3625 :     better if thread-safety is of any concern because they can be wrapped.
3626 :    
3627 :     *****************************************
3628 :    
3629 :     One bug fix: The function blockPlacement in three of the MLRISC
3630 :     backpatch files used to be hard-wired to one of two possibilities at
3631 :     link time (according to the value of the placementFlag). But (I
3632 :     think) it should rather sense the flag every time.
3633 :    
3634 :     *****************************************
3635 :    
3636 :     Other assorted changes (by other people who did not supply a HISTORY entry):
3637 :    
3638 :     1. the cross-module inliner now works much better (Monnier)
3639 :     2. representation of weights, frequencies, and probabilities in MLRISC
3640 :     changed in preparation of using those for weighted block placement
3641 :     (Reppy, George)
3642 :    
3643 :     ----------------------------------------------------------------------
3644 : george 1116 Name: Lal George
3645 : george 1124 Date: 2002/03/07 14:44:24 EST 2002
3646 :     Tag: george-20020307-weighted-block-placement
3647 :    
3648 :     Tested the weighted block placement optimization on all architectures
3649 :     (except the hppa) using AMPL to generate the block and edge frequencies.
3650 :     Changes were required in the machine properties to correctly
3651 : blume 2155 categorize trap instructions. There is an MLRISC flag
3652 :     "weighted-block-placement" that can be used to enable weighted block
3653 : george 1124 placement, but this will be ineffective without block/edge
3654 :     frequencies (coming soon).
3655 :    
3656 :    
3657 :     ----------------------------------------------------------------------
3658 :     Name: Lal George
3659 : blume 2155 Date: 2002/03/05 17:24:48 EST
3660 : george 1116 Tag: george-20020305-linkage-cluster
3661 :    
3662 : george 1124 In order to support the block placement optimization, a new cluster
3663 :     is generated as the very first cluster (called the linkage cluster).
3664 : blume 2155 It contains a single jump to the 'real' entry point for the compilation
3665 : george 1124 unit. Block placement has no effect on the linkage cluster itself, but
3666 :     all the other clusters have full freedom in the manner in which they
3667 : blume 2155 reorder blocks or functions.
3668 : george 1116
3669 :     On the x86 the typical linkage code that is generated is:
3670 :     ----------------------
3671 :     .align 2
3672 :     L0:
3673 :     addl $L1-L0, 72(%esp)
3674 : george 1124 jmp L1
3675 : george 1116
3676 :    
3677 :     .align 2
3678 :     L1:
3679 :     ----------------------
3680 :    
3681 : blume 2155 72(%esp) is the memory location for the stdlink register. This
3682 : george 1116 must contain the address of the CPS function being called. In the
3683 : george 1124 above example, it contains the address of L0; before
3684 : george 1116 calling L1 (the real entry point for the compilation unit), it
3685 :     must contain the address for L1, and hence
3686 :    
3687 :     addl $L1-L0, 72(%esp)
3688 :    
3689 : george 1124 I have tested this on all architectures except the hppa.The increase
3690 :     in code size is of course negligible
3691 : george 1116
3692 :     ----------------------------------------------------------------------
3693 : leunga 1115 Name: Allen Leung
3694 :     Date: 2002/03/03 13:20:00 EST
3695 :     Tag: leunga-20020303-mlrisc-tools
3696 :    
3697 :     Added #[ ... ] expressions to mlrisc tools
3698 :    
3699 :     ----------------------------------------------------------------------
3700 : blume 1096 Name: Matthias Blume
3701 : blume 1098 Date: 2002/02/27 12:29:00 EST
3702 :     Tag: blume-20020227-cdebug
3703 :     Description:
3704 :    
3705 :     - made types in structure C and C_Debug to be equal
3706 :     - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
3707 :     - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
3708 :    
3709 :     ----------------------------------------------------------------------
3710 :     Name: Matthias Blume
3711 : blume 1096 Date: 2002/02/26 12:00:00 EST
3712 :     Tag: blume-20020226-ffi
3713 :     Description:
3714 :    
3715 :     1. Fixed a minor bug in CM's "noweb" tool:
3716 :     If numbering is turned off, then truly don't number (i.e., do not
3717 :     supply the -L option to noweb). The previous behavior was to supply
3718 :     -L'' -- which caused noweb to use the "default" line numbering scheme.
3719 :     Thanks to Chris Richards for pointing this out (and supplying the fix).
3720 :    
3721 :     2. Once again, I reworked some aspects of the FFI:
3722 :    
3723 :     A. The incomplete/complete type business:
3724 :    
3725 :     - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
3726 :     gone!
3727 :     - ML types representing an incomplete type are now *equal* to
3728 :     ML types representing their corresponding complete types (just like
3729 :     in C). This is still safe because ml-nlffigen will not generate
3730 :     RTTI for incomplete types, nor will it generate functions that
3731 :     require access to such RTTI. But when ML code generated from both
3732 :     incomplete and complete versions of the C type meet, the ML types
3733 :     are trivially interoperable.
3734 :    
3735 :     NOTE: These changes restore the full generality of the translation
3736 :     (which was previously lost when I eliminated functorization)!
3737 :    
3738 :     B. Enum types:
3739 :    
3740 :     - Structure C now has a type constructor "enum" that is similar to
3741 :     how the "su" constructor works. However, "enum" is not a phantom
3742 :     type because each "T enum" has values (and is isomorphic to
3743 :     MLRep.Signed.int).
3744 :     - There are generic access operations for enum objects (using
3745 :     MLRep.Signed.int).
3746 :     - ml-nlffigen will generate a structure E_foo for each "enum foo".
3747 :     * The structure contains the definition of type "mlrep" (the ML-side
3748 :     representation type of the enum). Normally, mlrep is the same
3749 :     as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
3750 :     then mlrep will be defined as a datatype -- thus facilitating
3751 :     pattern matching on mlrep values.
3752 :     ("-ec" will be suppressed if there are duplicate values in an
3753 :     enumeration.)
3754 :     * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
3755 :     will be generated for each C enum constant xxx.
3756 :     * Conversion functions m2i and i2m convert between mlrep and
3757 :     MLRep.Signed.int. (Without "-ec", these functions are identities.)
3758 :     * Coversion functions c and ml convert between mlrep and "tag enum".
3759 :     * Access functions (get/set) fetch and store mlrep values.
3760 :     - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
3761 :     enumerations are merged into one single enumeration represented by
3762 :     structure E_'.
3763 :    
3764 :     ----------------------------------------------------------------------
3765 : leunga 1080 Name: Allen Leung
3766 : leunga 1094 Date: 2002/02/25 04:45:00 EST
3767 :     Tag: leunga-20020225-cps-spill
3768 :    
3769 :     This is a new implementation of the CPS spill phase.
3770 :     The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
3771 :     In case of problems, replace it with the old file spill.sml
3772 :    
3773 : blume 2155 The current compiler runs into some serious performance problems when
3774 : leunga 1094 constructing a large record. This can happen when we try to compile a
3775 :     structure with many items. Even a very simple structure like the following
3776 :     makes the compiler slow down.
3777 :    
3778 :     structure Foo = struct
3779 :     val x_1 = 0w1 : Word32.int
3780 :     val x_2 = 0w2 : Word32.int
3781 :     val x_3 = 0w3 : Word32.int
3782 :     ...
3783 :     val x_N = 0wN : Word32.int
3784 :     end
3785 :    
3786 : blume 2155 The following table shows the compile time, from N=1000 to N=4000,
3787 : leunga 1094 with the old compiler:
3788 :    
3789 :     N
3790 : blume 2155 1000 CPS 100 spill 0.04u 0.00s 0.00g
3791 :     MLRISC ra 0.06u 0.00s 0.05g
3792 : leunga 1094 (spills = 0 reloads = 0)
3793 : blume 2155 TOTAL 0.63u 0.07s 0.21g
3794 : leunga 1094
3795 : blume 2155 1100 CPS 100 spill 8.25u 0.32s 0.64g
3796 :     MLRISC ra 5.68u 0.59s 3.93g
3797 : leunga 1094 (spills = 0 reloads = 0)
3798 : blume 2155 TOTAL 14.71u 0.99s 4.81g
3799 : leunga 1094
3800 : blume 2155 1500 CPS 100 spill 58.55u 2.34s 1.74g
3801 :     MLRISC ra 5.54u 0.65s 3.91g
3802 : leunga 1094 (spills = 543 reloads = 1082)
3803 : blume 2155 TOTAL 65.40u 3.13s 6.00g
3804 : leunga 1094
3805 : blume 2155 2000 CPS 100 spill 126.69u 4.84s 3.08g
3806 :     MLRISC ra 0.80u 0.10s 0.55g
3807 : leunga 1094 (spills = 42 reloads = 84)
3808 : blume 2155 TOTAL 129.42u 5.10s 4.13g
3809 : leunga 1094
3810 : blume 2155 3000 CPS 100 spill 675.59u 19.03s 11.64g
3811 :     MLRISC ra 2.69u 0.27s 1.38g
3812 : leunga 1094 (spills = 62 reloads = 124)
3813 : blume 2155 TOTAL 682.48u 19.61s 13.99g
3814 : leunga 1094
3815 : blume 2155 4000 CPS 100 spill 2362.82u 56.28s 43.60g
3816 :     MLRISC ra 4.96u 0.27s 2.72g
3817 : leunga 1094 (spills = 85 reloads = 170)
3818 : blume 2155 TOTAL 2375.26u 57.21s 48.00g
3819 : leunga 1094
3820 : blume 2155 As you can see the old cps spill module suffers from some serious
3821 :     performance problem. But since I cannot decipher the old code fully,
3822 :     instead of patching the problems up, I'm reimplementing it
3823 :     with a different algorithm. The new code is more modular,
3824 :     smaller when compiled, and substantially faster
3825 :     (O(n log n) time and O(n) space). Timing of the new spill module:
3826 : leunga 1094
3827 : blume 2155 4000 CPS 100 spill 0.02u 0.00s 0.00g
3828 :     MLRISC ra 0.25u 0.02s 0.15g
3829 :     (spills=1 reloads=3)
3830 :     TOTAL 7.74u 0.34s 1.62g
3831 : leunga 1094
3832 :     Implementation details:
3833 :    
3834 : blume 2155 As far as I can tell, the purpose of the CPS spill module is to make sure the
3835 :     number of live variables at any program point (the bandwidth)
3836 :     does not exceed a certain limit, which is determined by the
3837 :     size of the spill area.
3838 : leunga 1094
3839 : blume 2155 When the bandwidth is too large, we decrease the register pressure by
3840 : leunga 1094 packing live variables into spill records. How we achieve this is
3841 :     completely different than what we did in the old code.
3842 :    
3843 : blume 2155 First, there is something about the MLRiscGen code generator
3844 : leunga 1094 that we should be aware of:
3845 :    
3846 :     o MLRiscGen performs code motion!
3847 : blume 2155
3848 : leunga 1094 In particular, it will move floating point computations and
3849 : blume 2155 address computations involving only the heap pointer to
3850 :     their use sites (if there is only a single use).
3851 : leunga 1094 What this means is that if we have a CPS record construction
3852 :     statement
3853 : blume 2155
3854 : leunga 1094 RECORD(k,vl,w,e)
3855 : blume 2155
3856 :     we should never count the new record address w as live if w
3857 : leunga 1094 has only one use (which is often the case).
3858 : blume 2155
3859 : leunga 1094 We should do something similar to floating point, but the transformation
3860 :     there is much more complex, so I won't deal with that.
3861 :    
3862 :     Secondly, there are now two new cps primops at our disposal:
3863 :    
3864 : blume 2155 1. rawrecord of record_kind option
3865 : leunga 1094 This pure operator allocates some uninitialized storage from the heap.
3866 :     There are two forms:
3867 : blume 2155
3868 : leunga 1094 rawrecord NONE [INT n] allocates a tagless record of length n
3869 :     rawrecord (SOME rk) [INT n] allocates a tagged record of length n
3870 :     and initializes the tag.
3871 : blume 2155
3872 : leunga 1094 2. rawupdate of cty
3873 : blume 2155 rawupdate cty (v,i,x)
3874 : leunga 1094 Assigns to x to the ith component of record v.
3875 :     The storelist is not updated.
3876 : blume 2155
3877 : leunga 1094 We use these new primops for both spilling and increment record construction.
3878 : blume 2155
3879 : leunga 1094 1. Spilling.
3880 : blume 2155
3881 : leunga 1094 This is implemented with a linear scan algorithm (but generalized
3882 :     to trees). The algorithm will create a single spill record at the
3883 :     beginning of the cps function and use rawupdate to spill to it,
3884 :     and SELECT or SELp to reload from it. So both spills and reloads
3885 : blume 2155 are fine-grain operations. In contrast, in the old algorithm
3886 :     "spills" have to be bundled together in records.
3887 :    
3888 : leunga 1094 Ideally, we should sink the spill record construction to where
3889 :     it is needed. We can even split the spill record into multiple ones
3890 :     at the places where they are needed. But CPS is not a good
3891 :     representation for global code motion, so I'll keep it simple and
3892 :     am not attempting this.
3893 : blume 2155
3894 : leunga 1094 2. Incremental record construction (aka record splitting).
3895 :    
3896 :     Long records with many component values which are simulatenously live
3897 : blume 2155 (recall that single use record addresses are not considered to
3898 : leunga 1094 be live) are constructed with rawrecord and rawupdate.
3899 :     We allocate space on the heap with rawrecord first, then gradually
3900 :     fill it in with rawupdate. This is the technique suggested to me
3901 :     by Matthias.
3902 : blume 2155
3903 : leunga 1094 Some restrictions on when this is applicable:
3904 :     1. It is not a VECTOR record. The code generator currently does not handle
3905 :     this case. VECTOR record uses double indirection like arrays.
3906 : blume 2155 2. All the record component values are defined in the same "basic block"
3907 :     as the record constructor. This is to prevent speculative
3908 :     record construction.
3909 : leunga 1094
3910 :     ----------------------------------------------------------------------
3911 :     Name: Allen Leung
3912 : leunga 1086 Date: 2002/02/22 01:02:00 EST
3913 :     Tag: leunga-20020222-mlrisc-tools
3914 :    
3915 :     Minor bug fixes in the parser and rewriter
3916 :    
3917 :     ----------------------------------------------------------------------
3918 :     Name: Allen Leung
3919 : leunga 1085 Date: 2002/02/21 20:20:00 EST
3920 :     Tag: leunga-20020221-peephole
3921 :    
3922 :     Regenerated the peephole files. Some contained typos in the specification
3923 :     and some didn't compile because of pretty printing bugs in the old version
3924 :     of 'nowhere'.
3925 :    
3926 :     ----------------------------------------------------------------------
3927 :     Name: Allen Leung
3928 : leunga 1079 Date: 2002/02/19 20:20:00 EST
3929 :     Tag: leunga-20020219-mlrisc-tools
3930 :     Description:
3931 :    
3932 :     Minor bug fixes to the mlrisc-tools library:
3933 :    
3934 :     1. Fixed up parsing colon suffixed keywords
3935 :     2. Added the ability to shut the error messages up
3936 :     3. Reimplemented the pretty printer and fixed up/improved
3937 :     the pretty printing of handle and -> types.
3938 :     4. Fixed up generation of literal symbols in the nowhere tool.
3939 :     5. Added some SML keywords to to sml.sty
3940 :    
3941 :     ----------------------------------------------------------------------
3942 :     Name: Matthias Blume
3943 : blume 1078 Date: 2002/02/19 16:20:00 EST
3944 :     Tag: blume-20020219-cmffi
3945 :     Description:
3946 :    
3947 :     A wild mix of changes, some minor, some major:
3948 :    
3949 :     * All C FFI-related libraries are now anchored under $c:
3950 :     $/c.cm --> $c/c.cm
3951 :     $/c-int.cm --> $c/internals/c-int.cm
3952 :     $/memory.cm --> $c/memory/memory.cm
3953 :    
3954 :     * "make" tool (in CM) now treats its argument pathname slightly
3955 :     differently:
3956 :     1. If the native expansion is an absolute name, then before invoking
3957 :     the "make" command on it, CM will apply OS.Path.mkRelative
3958 :     (with relativeTo = OS.FileSys.getDir()) to it.
3959 :     2. The argument will be passed through to subsequent phases of CM
3960 :     processing without "going native". In particular, if the argument
3961 :     was an anchored path, then "make" will not lose track of that anchor.
3962 :    
3963 :     * Compiler backends now "know" their respective C calling conventions
3964 :     instead of having to be told about it by ml-nlffigen. This relieves
3965 :     ml-nlffigen from one of its burdens.
3966 :    
3967 :     * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3968 :    
3969 :     * Export C_DEBUG and C_Debug from $c/c.cm.
3970 :    
3971 :     * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3972 :     subtyping relationship between incomplete pointers and their complete
3973 :     counterparts. For this, ('t, 'c) ptr has been changed to 'o ptr --
3974 :     with the convention of instantiating 'o with ('t, 'c) obj whenever
3975 :     the pointer target type is complete. In the incomplete case, 'o
3976 :     will be instantiated with some "'c iobj" -- a type obtained by
3977 :     using one of the functors PointerToIncompleteType or PointerToCompleteType.
3978 :    
3979 :     Operations that work on both incomplete and complete pointer types are
3980 :     typed as taking an 'o ptr while operations that require the target to
3981 :     be known are typed as taking some ('t, 'c) obj ptr.
3982 :    
3983 :     voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3984 :     where void is an eqtype without any values. This makes it possible
3985 :     to work on voidptr values using functions meant to operate on light
3986 :     incomplete pointers.
3987 :    
3988 :     * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3989 :     been vastly simplified.
3990 :    
3991 :     ----------------------------------------------------------------------
3992 :     Name: Matthias Blume
3993 : blume 1076 Date: 2002/02/19 10:48:00 EST
3994 :     Tag: blume-20020219-pqfix
3995 :     Description:
3996 :    
3997 :     Applied Chris Okasaki's bug fix for priority queues.
3998 :    
3999 :     ----------------------------------------------------------------------
4000 :     Name: Matthias Blume
4001 : blume 1073 Date: 2002/02/15 17:05:00
4002 : blume 1069 Tag: Release_110_39
4003 :     Description:
4004 :    
4005 : blume 1073 Last-minute retagging is becoming a tradition... :-(
4006 :    
4007 :     This is the working release 110.39.
4008 :    
4009 :     ----------------------------------------------------------------------
4010 :     Name: Matthias Blume
4011 :     Date: 2002/02/15 16:00:00 EST
4012 :     Tag: Release_110_39-orig
4013 :     Description:
4014 :    
4015 : blume 1069 Working release 110.39. New bootfiles.
4016 :    
4017 : blume 1073 (Update: There was a small bug in the installer so it wouldn't work
4018 :     with all shells. So I retagged. -Matthias)
4019 :    
4020 : blume 1069 ----------------------------------------------------------------------
4021 :     Name: Matthias Blume
4022 : blume 1068 Date: 2002/02/15 14:17:00 EST
4023 :     Tag: blume-20020215-showbindings
4024 :     Description:
4025 :    
4026 :     Added EnvRef.listBoundSymbols and CM.State.showBindings. Especially
4027 :     the latter can be useful for exploring what bindings are available at
4028 :     the interactive prompt. (The first function returns only the list
4029 :     of symbols that are really bound, the second prints those but also the
4030 :     ones that CM's autoloading mechanism knows about.)
4031 :    
4032 :     ----------------------------------------------------------------------
4033 :     Name: Matthias Blume
4034 : blume 1067 Date: 2002/02/15 12:08:00 EST
4035 :     Tag: blume-20020215-iptrs
4036 :     Description:
4037 :    
4038 :     Two improvements to ml-nlffigen:
4039 :    
4040 :     1. Write files only if they do not exist or if their current contents
4041 :     do not coincide with what's being written. (That is, avoid messing
4042 :     with the time stamps unless absolutely necessary.)
4043 :    
4044 :     2. Implement a "repository" mechanism for generated files related
4045 :     to "incomplete pointer types". See the README file for details.
4046 :    
4047 :     ----------------------------------------------------------------------
4048 :     Name: Matthias Blume
4049 : blume 1066 Date: 2002/02/14 11:50:00 EST
4050 :     Tag: blume-20020214-quote
4051 :     Description:
4052 :    
4053 :     Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm). This is required
4054 :     because of the new and improved tag generation scheme. (Thanks to Allen
4055 :     Leung for pointing it out.)
4056 :    
4057 :     ----------------------------------------------------------------------
4058 : george 1065 Name: Lal George
4059 :     Date: 2002/02/14 09:55:27 EST 2002
4060 :     Tag: george-20020214-isabelle-bug
4061 : blume 2155 Description:
4062 : george 1065
4063 :     Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
4064 :     of Isabelle on the x86.
4065 :    
4066 :     From Allen:
4067 :     -----------
4068 : blume 2155 I've found the problem:
4069 : george 1065
4070 :     in ra-core.sml, I use the counter "blocked" to keep track of the
4071 :     true number of elements in the freeze queue. When the counter goes
4072 :     to zero, I skip examining the queue. But I've messed up the
4073 :     bookkeeping in combine():
4074 :    
4075 :     else ();
4076 :     case !ucol of
4077 : blume 2155 PSEUDO => (if !cntv > 0 then
4078 :     (if !cntu > 0 then blocked := !blocked - 1 else ();
4079 : george 1065 ^^^^^^^^^^^^^^^^^^^^^^^
4080 :     moveu := mergeMoveList(!movev, !moveu)
4081 :     )
4082 : blume 2155 else ();
4083 :    
4084 : george 1065 combine() is called to coalesce two nodes u and v.
4085 :     I think I was thinking that if the move counts of u and v are both
4086 :     greater than zero then after they are coalesced then one node is
4087 :     removed from the freeze queue. Apparently I was thinking that
4088 :     both u and v are of low degree, but that's clearly not necessarily true.
4089 :    
4090 : blume 2155
4091 : george 1065 02/12/2002:
4092 :     Here's the patch. HOL now compiles.
4093 :    
4094 :     I don't know how this impact on performance (compile
4095 : blume 2155 time or runtime). This bug caused the RA (especially on the x86)
4096 : george 1065 to go thru the potential spill phase when there are still nodes on the
4097 :     freeze queue.
4098 :    
4099 :    
4100 :    
4101 :    
4102 :     ----------------------------------------------------------------------
4103 : blume 1058 Name: Matthias Blume
4104 : blume 1064 Date: 2002/02/13 22:40:00 EST
4105 :     Tag: blume-20020213-fptr-rtti
4106 :     Description:
4107 :    
4108 :     Fixed a bug in ml-nlffigen that was introduced with one of the previous
4109 :     updates.
4110 :    
4111 :     ----------------------------------------------------------------------
4112 :     Name: Matthias Blume
4113 : blume 1063 Date: 2002/02/13 16:41:00 EST
4114 :     Tag: blume-20020213-cmlpq
4115 :     Description:
4116 :    
4117 :     Added new priority queue export symbols (which have just been added to
4118 :     smlnj-lib.cm) to CML's version of smlnj-lib.cm. (Otherwise CML would
4119 :     not compile and the installer would choke.)
4120 :    
4121 :     ----------------------------------------------------------------------
4122 :     Name: Matthias Blume
4123 : blume 1062 Date: 2002/02/13 16:15:00 EST
4124 :     Tag: blume-20020213-various
4125 :     Description:
4126 :    
4127 :     1. More tweaks to ml-nlffigen:
4128 :    
4129 :     - better internal datastructures (resulting in slight speedup)
4130 :     - "-match" option requires exact match
4131 :     - "localized" gensym counters (untagged structs/unions nested within
4132 :     other structs/unions or within typedefs get a fresh counter; their
4133 :     tag will be prefixed by a concatenation of their parents' tags)
4134 :     - bug fixes (related to calculation of transitive closure of types
4135 :     to be included in the output)
4136 :    
4137 :     2. Minor Basis updates:
4138 :    
4139 :     - added implementations for List.collate and Option.app
4140 :    
4141 :     ----------------------------------------------------------------------
4142 :     Name: Matthias Blume
4143 : blume 1060 Date: 2002/02/11 15:55:00 EST
4144 :     Tag: blume-20020211-gensym
4145 :     Description:
4146 :    
4147 :     Added a "-gensym" option to command line of ml-nlffigen. This can be
4148 :     used to specify a "stem" -- a string that is inserted in all "gensym'd"
4149 :     names (ML structure names that correspond to unnamed C structs, unions,
4150 :     and enums), so that separate runs of ml-nlffigen do not clash.
4151 :    
4152 :     ----------------------------------------------------------------------
4153 :     Name: Matthias Blume
4154 : blume 1059 Date: 2002/02/11 12:05:00 EST
4155 :     Tag: blume-20020211-gensml
4156 :     Description:
4157 :