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

SCM Repository

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

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

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 890, Thu Jul 19 20:38:56 2001 UTC revision 1898, Mon Feb 27 05:25:47 2006 UTC
# Line 8  Line 8 
8  The form of an entry should be:  The form of an entry should be:
9    
10  Name:  Name:
11  Date:  Date: yyyy/mm/dd
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: Matthias Blume (http://www.tti-c.org/blume/)
17    Date: 2006/02/26 23:30:00 CST
18    Tag: blume-20060226-ml-flex
19    Description:
20    
21    Removed ml-flex' dependency on regexp-lib.cm.  Turned installation
22    of ml-flex on by default.
23    
24    ----------------------------------------------------------------------
25    Name: Matthias Blume (http://www.tti-c.org/blume/)
26    Date: 2006/02/24 16:30:00 CST
27    Tag: blume-20060224-ml-flex
28    Description:
29    
30    Added ml-flex sources.  Partially integrated, but should not be
31    turned on yet!  (Read: leave it commented-out in config/targets!)
32    
33    ----------------------------------------------------------------------
34    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
35    Date: 2006/02/23
36    Tag: jhr-20060223-x86-64
37    Description:
38    Changes to support building on x86-64 systems (using the 32-bit mode).
39    Also cleaned up signal handling on Linux.  Support for pre-2.2 Linux
40    kernels dropped.
41    
42    ----------------------------------------------------------------------
43    Name: Allen Leung (leunga (at) reservoir (dot) com)
44    Date: 2006/02/22 15:07:00 EST
45    Tag: leunga-20060222-mdl-gen
46    Description:
47    
48      CM has changed.  Updated the script for rebuilding the MLRISC generated
49    files.
50    
51    ----------------------------------------------------------------------
52    Name: Matthias Blume (http://www.tti-c.org/blume/)
53    Date: 2006/02/14 17:30:00 CST
54    Tag: blume-20060214-intel-mac
55    Description:
56    
57    Hooked code for Darwin-specific Intel ABI into main compiler.
58    (This is a temporary solution which relies on the fact that the
59    compiler itself does not use NLFFI.  Eventually we need to divorce
60    intel mac from generic x86 unix code and make separate sets of binaries.)
61    
62    ----------------------------------------------------------------------
63    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
64    Date: 2006/02/13
65    Tag: jhr-20060213-intel-mac
66    Description:
67    Changed MLRISC x86 CCalls for partial support of MacOS X ABI.
68    
69    ----------------------------------------------------------------------
70    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
71    Date: 2006/02/06
72    Tag: jhr-20060206-intel-mac
73    Description:
74    Changes to support MacOS X on Intel hardware.  The C-calls support in
75    MLRISC must also be updated to support the MacOS X ABI.
76    
77    ----------------------------------------------------------------------
78    Name: Matthias Blume (http://www.tti-c.org/blume/)
79    Date: 2005/12/16 00:15:00 CST
80    Tag: blume-20051216-cm
81    Description:
82    
83    Improved error reporting and handling in CM.
84    
85    ----------------------------------------------------------------------
86    Name: Matthias Blume (http://www.tti-c.org/blume/)
87    Date: 2005/11/21 14:00:00 CST
88    Tag: blume-20051121-html-lib
89    Description:
90    
91    Enabling $/html-lib.cm in config/preloads did not work.  This is fixed
92    now.  (Since the anchor mapping for html-lib.cm is not yet in effect
93    at the time when config/preloads is processed, the library has to
94    be referred to by another name.  In 110.57 this would be
95    $SMLNJ-LIB/HTML/html-lib.cm.  I arranged for $smlnj/smlnj-lib/html-lib.cm
96    to be valid as well.)
97    
98    Thanks to Todd Wilson (Fresno) for alerting me to this issue.
99    
100    ----------------------------------------------------------------------
101    Name: Matthias Blume (http://www.tti-c.org/blume/)
102    Date: 2005/11/19 13:45:00 CST
103    Tag: Release_110_57
104    Description:
105    
106    Moved release tag for 110.57.
107    
108    ----------------------------------------------------------------------
109    Name: Matthias Blume (http://www.tti-c.org/blume/)
110    Date: 2005/11/19 23:35:00 CST
111    Tag: blume-20051119-fish
112    Description:
113    
114    Fixed a problem in config/install.sh where it tries to "fish" the
115    name of the CM metadata directory from the wrong place (because
116    the physical location of basis.cm has changed).  Also, corrected the
117    path anchor for $/html-lib.cm.  (Thanks to M. Fluet for pointing out
118    these problems.)
119    
120    ----------------------------------------------------------------------
121    Name: Matthias Blume (http://www.tti-c.org/blume/)
122    Date: 2005/11/16 14:30:00 CST
123    Tag: blume-orig-110_57
124    Description:
125    
126    New working version (110.57).  NEW BOOTFILES!
127    
128    UPDATE: The release tag has been moved to include a fix for a brown-paper-
129            bag glitch (see above).
130    
131    ----------------------------------------------------------------------
132    Name: Matthias Blume (http://www.tti-c.org/blume/)
133    Date: 2005/11/16 10:25:00 CST
134    Tag: blume-22051116-btrace
135    Description:
136    
137    Fixed problem with bogus exception message when using back-trace
138    facility.
139    
140    ----------------------------------------------------------------------
141    Name: Matthias Blume (http://www.tti-c.org/blume/)
142    Date: 2005/11/15 22:35:00 CST
143    Tag: blume-22051115-array2copy
144    Description:
145    
146    Added simple implementation of Array2.copy.  (Warning: mostly untested.)
147    
148    ----------------------------------------------------------------------
149    Name: Matthias Blume (http://www.tti-c.org/blume/)
150    Date: 2005/11/15 21:35:00 CST
151    Tag: blume-22051115-versiontool
152    Description:
153    
154    Reversed change to src/system/smlnj/internals/versiontool.cm.  This
155    file gets loaded as a tool -- by the equivalent of CM.make during the
156    run of CMB.make.  Thus, CMB's path configuration is meaningless for
157    it.  Instead, it has the status of "user code", so it should use
158    $/basis.cm to refer to the Basis library.  (At least that's true for
159    the purpose of bootstrapping the previous change.  In the future it
160    might make sense to have versiontool.cm refer to
161    $smlnj/basis/basis.cm, i.e., the version of the Basis that the
162    compiler itself uses.)
163    
164    Also patched src/system/testml to have it activate those extra anchor
165    bindings in config/extrapathconfig.
166    
167    ----------------------------------------------------------------------
168    Name: Matthias Blume (http://www.tti-c.org/blume/)
169    Date: 2005/11/15 16:35:00 CST
170    Tag: blume-22051115-libraries
171    Description:
172    
173    This change affects the way the following libraries are tied into
174    the system:
175    
176       $/basis.cm
177       $/smlnj-lib.cm
178       $/pp-lib.cm
179       $/controls-lib.cm
180       $/html-lib.cm
181       $/ml-yacc-lib.cm
182    
183    These libraries are now internally (as seen from the source code
184    of the implementation itself) known by the following names:
185    
186       $smlnj/basis/basis.cm
187       $smlnj/smlnj-lib/smlnj-lib.cm
188       $smlnj/smlnj-lib/pp-lib.cm
189       $smlnj/smlnj-lib/controls-lib.cm
190       $smlnj/smlnj-lib/html-lib.cm
191       $smlnj/ml-yacc/ml-yacc-lib.cm
192    
193    This makes it possible to work with code that requires different
194    versions of these libraries, and which refers to these libraries using
195    their "default" names (i.e., the first set of names above).  In other
196    words, one can un-define or re-define those default names without
197    compromising the proper functioning of the compiler itself.
198    
199    A similar procedure had already been performed for several of the
200    MLRISC libraries that are linked into the compiler.  I did some
201    cleanup on this code.
202    
203    A new file in the config directory (named extrapathconfig) is
204    responsible for setting up path anchors that the compiler itself does
205    not need, but that are typically required by user code.
206    
207    ----------------------------------------------------------------------
208    Name: Matthias Blume
209    Date: 2005/11/07 17:05:00 CST
210    Tag: blume-20051107-slice
211    Description:
212    
213    Fixed erroneous out-of-bounds test in the "update" function of
214    various *ArraySlice modules.  (Thanks to Vesa A. Norrman for pointing
215    out the problem.)
216    
217    Pushed some Basis changes through ML-Lex, CML, and eXene.
218    
219    ----------------------------------------------------------------------
220    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
221    Date: 2005/11/07
222    Description:
223    Fixed a Basis incompatibility: The depreciated function Substring.all
224    was removed (use Substring.full instead).
225    
226    ----------------------------------------------------------------------
227    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
228    Date: 2005/11/05
229    Description:
230    Tweaked interval set API in SML/NJ library; see the CHANGES file for details.
231    
232    ----------------------------------------------------------------------
233    Name: Matthias Blume (http://www.tti-c.org/blume/)
234    Date: 2005/11/02 16:25:00 CST
235    Tag: blume-20051102-hex
236    Description:
237    
238    * Runtime system bootstrap code now accepts hex digits in BOOTLIST
239      in either upper case or lower case format.
240    * Pushed changes to names of Pack<N>{Big,Little} structures through
241      CML and eXene.
242    
243    ----------------------------------------------------------------------
244    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
245    Date: 2005/11/02
246    Description:
247    Fixed a Basis incompatibility: Pack<N>{Big,Little} structures should
248    be called PackWord<N>{Big,Little}.
249    
250    ----------------------------------------------------------------------
251    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
252    Date: 2005/10/28
253    Description:
254    Fixed a minor Basis incompatibility: hex digits should be upper case.
255    
256    ----------------------------------------------------------------------
257    Name: Matthias Blume (blume (at) tti - c (dot) org)
258    Date: 2005/10/25 16:00:00 CDT
259    Tag: Release_110_56
260    Description:
261    
262    New working version (110.56).  NEW BOOTFILES!
263    
264    ----------------------------------------------------------------------
265    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
266    Date: 2005/10/25
267    Description:
268    Added interval sets to utility library (signatures INTERVAL_DOMAIN
269    and INTERVAL_SET, and functor IntervalSetFn).
270    
271    ----------------------------------------------------------------------
272    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
273    Date: 2005/10/14
274    Description:
275    Add Zhong Shao's fix for datatype equality functions.
276    
277    ----------------------------------------------------------------------
278    Name: Allen Leung (leunga (at) reservoir (dot) com)
279    Date: 2005/10/14 12:30:00 EST
280    Tag: leunga-20051014-x86-spill
281    Description:
282    
283    Bug fix a bug found by Carl Hauser.
284    There was a typo in the reload code for FCMP in x86SpillInstr.sml.
285    
286    ----------------------------------------------------------------------
287    Name: Allen Leung (leunga (at) reservoir (dot) com)
288    Date: 2005/10/14 11:15:00 EST
289    Tag: leunga-20051014-x86-asm
290    Description:
291    
292    Removed some debugging code in file x86Asm.sml.
293    The function emit_operand was printing out debugging output.
294    
295    ----------------------------------------------------------------------
296    Name: John Reppy
297    Date: 2005/07/27
298    Tag:
299    Description:
300    
301    Fixed ml-lex to recognize "\r" as representing carriage return.
302    
303    ----------------------------------------------------------------------
304    Name: John Reppy
305    Date: 2005/07/27
306    Tag:
307    Description:
308    
309    Fixed ml-yacc to work on files with non-native end-of-line
310    encodings (e.g., Windows text file on a Unix system).
311    
312    ----------------------------------------------------------------------
313    Name: John Reppy
314    Date: 2005/07/20
315    Tag:
316    Description:
317    
318    Added changes from Dominic Evans (oldmanuk (at) gmail (dot) com)
319    to support HPUX 11.
320    
321    ----------------------------------------------------------------------
322    Name: John Reppy
323    Date: 2005/07/06
324    Tag:
325    Description:
326    
327    Changes to the SML/NJ library.  See smlnj-lib/CHANGES for details.
328    
329    ----------------------------------------------------------------------
330    Name: Matthias Blume (blume (at) tti - c (dot) org)
331    Date: 2005/07/06 00:45:00 CDT
332    Tag: blume-20050706-slice-copy
333    Description:
334    
335    Fixed reversed logic for deciding whether to "copy up" or "copy down"
336    in *-array-slice.sml.
337    
338    ----------------------------------------------------------------------
339    Name: Allen Leung (leunga (at) reservoir (dot) com)
340    Date: 2005/05/31 17:00:00 EST
341    Tag: leunga-20050531-cygwin-fault-2
342    Description:
343    
344    A typo in the cygwin code fixed.
345    
346    ----------------------------------------------------------------------
347    Name: Allen Leung (leunga (at) reservoir (dot) com)
348    Date: 2005/05/31 16:47:00 EST
349    Tag: leunga-20050531-cygwin-fault
350    Description:
351    
352    Updated Cygwin's fault/signal handling to match the Windows version.
353    Updated the export list.
354    
355    ----------------------------------------------------------------------
356    Name: Matthias Blume (blume (at) tti - c (dot) org)
357    Date: 2005/05/18 13:35:00 CDT
358    Tag: Release_110_54
359    Description:
360    
361    New working version (110.54).  NEW BOOTFILES!
362    
363    ----------------------------------------------------------------------
364    Name: Matthias Blume (blume (at) tti - c (dot) org)
365    Date: 2005/05/18 11:58:00 CDT
366    Tag: blume-20050518-installer
367    Description:
368    
369    Added support scripts for Mac OS X PackageMaker and modified
370    config/install.sh so that it supports re-dumping a heap image
371    after customization.
372    
373    ----------------------------------------------------------------------
374    Name: Matthias Blume (blume (at) tti - c (dot) org)
375    Date: 2005/05/18 10:55:00 CDT
376    Tag: blume-20050518-realdiv-noovld
377    Description:
378    
379    Un-overloaded / to work around bug in overloading resolution code.
380    
381    ----------------------------------------------------------------------
382    Name: Matthias Blume (blume (at) tti - c (dot) org)
383    Date: 2005/05/16 23:50:00 CDT
384    Tag: blume-20050516-redump-heap
385    Description:
386    
387    Added mechanism for re-creating a heap file for the interactive system
388    after configuration variables have been changed.
389    
390       CM.redump_heap : string -> unit
391    
392    This is much like SMLofNJ.exportML, but starting from the resulting
393    heap does not return to the caller of CM.redump_heap but
394    restarts the interactive system from scratch.  The original call of
395    CM.redump_heap does not return but ends the interactive session.
396    Thus, CM.redump_heap is a lot like SMLofNJ.exportFn.
397    
398    Internally, redump_heap winds the dynamic execution context back to
399    the point where the original heap image was created and re-executes
400    the heap image generation code in the boot code.
401    
402    ----------------------------------------------------------------------
403    Name: Matthias Blume (blume (at) tti - c (dot) org)
404    Date: 2005/05/09 21:55:00 CDT
405    Tag: blume-20050509-word64
406    Description:
407    
408    Added a hack to the existing hack known as Word64 to make fromString
409    behave correctly.  I am still not sure whether Word64.scan will work
410    as specified with respect to the interaction of radix and prefix.
411    
412    ----------------------------------------------------------------------
413    Name: Allen Leung (leunga (at) reservoir (dot) com)
414    Date: 2005/05/04 11:50:00 EST
415    Tag: leunga-20050504-checkgc
416    Description:
417    
418    Added a gc protocol checking phase.  This phase is enabled with
419    the flag "check-gc".  "debug-check-gc" turns on the verbose mode.
420    
421    ----------------------------------------------------------------------
422    Name: Matthias Blume (blume (at) tti - c (dot) org)
423    Date: 2005/05/04 10:45:00 CDT
424    Tag: blume-20050504-intinf
425    Description:
426    
427    Fixed a bug in the implementation of div and mod for IntInf.
428    Thanks to Neophytos Michael for reporting the problem.
429    
430    ----------------------------------------------------------------------
431    Name: Matthias Blume (blume (at) tti - c (dot) org)
432    Date: 2005/05/04 10:35:00 CDT
433    Tag: blume-20050504-join
434    Description:
435    
436    Added a "join" combinator to the ParserComb module in smlnj-lib.cm.
437    
438    ----------------------------------------------------------------------
439    Name: Matthias Blume (blume (at) tti - c (dot) org)
440    Date: 2005/02/28 23:40:00 CST
441    Tag: blume-20050228-mVar
442    Description:
443    
444    Fixed serious bug (brown paper bag variety) in new implementation of
445    structure Atom in CML.  (I had accidentally used a mailbox instead of
446    an mvar, leaving the door open for races.)
447    
448    ----------------------------------------------------------------------
449    Name: Matthias Blume (blume (at) tti - c (dot) org)
450    Date: 2005/02/25 15:00:00 CST
451    Tag: Release_110_53
452    Description:
453    
454    New working version (110.53).  NEW BOOTFILES!
455    
456    ----------------------------------------------------------------------
457    Name: Matthias Blume (blume (at) tti - c (dot) org)
458    Date: 2005/02/25 14:50:00 CST
459    Tag: blume-20050225-susp
460    Description:
461    
462    Brought back SMLofNJ.Susp.  The underlying suspension type is the one
463    implemented in Core -- which means that it is the same as the one used
464    by the lazy extension.
465    
466    ----------------------------------------------------------------------
467    Name: Matthias Blume (blume (at) tti - c (dot) org)
468    Date: 2005/02/24 16:50:00 CST
469    Tag: blume-20050224-cml-atom
470    Description:
471    
472    Simpler and at the same time more general implementation of structure
473    Atom in CML.
474    
475    ----------------------------------------------------------------------
476    Name: Matthias Blume (blume (at) tti - c (dot) org)
477    Date: 2005/02/15 17:35:00 CST
478    Tag: blume-20050215-tools
479    Description:
480    
481    Created new "tools" directory under "src" and moved "TraceDebugProf"
482    there.
483    
484    ----------------------------------------------------------------------
485    Name: Matthias Blume (blume (at) tti - c (dot) org)
486    Date: 2005/02/10 17:55:00 CST
487    Tag: blume-20050210-longlong
488    Description:
489    
490    Implemented "long long" arguments and results for NLFFI.  (Only the
491    PPC/MacOS implementation is complete, the other backends still need to
492    be updated.)
493    
494    ----------------------------------------------------------------------
495    Name: Matthias Blume (blume (at) tti - c (dot) org)
496    Date: 2005/01/24 17:40:00 CST
497    Tag: blume-20050124-mlyacc
498    Description:
499    
500    Minor cleanup in ML-Yacc rule printing mechanism.  This should fix a
501    problem with certain "as" patterns which previously got rendered
502    using incorrect syntax.
503    
504    ----------------------------------------------------------------------
505    Name: Matthias Blume (blume (at) tti - c (dot) org)
506    Date: 2005/01/18 12:00:00 CST
507    Tag: blume-20050118-profile
508    Description:
509    
510    Made time profiling code (interrupt handler) in runtime system aware
511    of new array representation.
512    
513    ----------------------------------------------------------------------
514    Name: Matthias Blume (blume (at) tti - c (dot) org)
515    Date: 2005/01/14 18:00:00 CST
516    Tag: blume-20050114-heap2exec
517    Description:
518    
519    Implemented new (but still experimental) heap2exec facility.  This is
520    tested under Mac OS X and should work under Linux (will test shortly).
521    It will probably also work on the Sparc (will test some time later).
522    
523      - removed old "HACKED_STANDALONE" hack from runtime
524    
525    To be able to test this, uncomment the request for "heap2asm" in
526    config/targets prior to installation.  (Notice that this is different
527    from "heap2exec" mentioned below.  Not a typo.)
528    
529    To perform an actual test, run the command
530    
531       $ bin/heap2exec heapfile execfile
532    
533    (You can put heap2exec on your shell's path.)
534    
535    For example, run
536    
537      $ bin/heap2exec bin/.heap/ml-yacc.ppc-darwin mly
538    
539    This will create a standalone executable called "mly" which you can
540    then invoke directly as a command.
541    
542    ----------------------------------------------------------------------
543    Name: Matthias Blume (blume (at) tti - c (dot) org)
544    Date: 2005/01/07 11:44:00 CST
545    Tag: blume-20050107-mlstring
546    Description:
547    
548    fixed off-by-one error in ML_STRING macro (globals.c)
549    
550    ----------------------------------------------------------------------
551    Name: Matthias Blume (blume (at) tti - c (dot) org)
552    Date: 2004/12/23 18:00:00 CST
553    Tag: blume-20041223-santa
554    Description:
555    
556    Made ml-build script "smarter" (but only very little).
557    
558    ----------------------------------------------------------------------
559    Name: Matthias Blume (blume (at) tti - c (dot) org)
560    Date: 2004/12/21 15:05:00 CST
561    Tag: blume-20041221-longlong
562    Description:
563    
564    * Implemented access to signed and unsigned long long data in NLFFI.
565    (The parameter-passing part of the picture has not complete. But
566    data structure access seems to work.)
567    
568    * Fixed CM's incorrect assumption that the PPC is little-endian.
569      (On the Mac, it is big-endian.  And that's currently our only
570      PPC platform.)
571    
572    ----------------------------------------------------------------------
573    Name: Matthias Blume (blume (at) tti - c (dot) org)
574    Date: 2004/12/21 12:50:00 CST
575    Tag: blume-20041221-memory
576    Description:
577    
578    Some cleanup in the $c/memory.cm library: separated some concerns by
579    moving allocation code and memory access code each into their own
580    files.
581    
582    ----------------------------------------------------------------------
583    Name: Allen Leung (leunga (at) reservoir (dot) com)
584    Date: 2004/12/17 16:12:00 EST
585    Tag: leunga-20041217-cygwin-smlnj-home
586    Description:
587    
588        The Unix I/O library of SML/NJ on cygwin does not understand
589    Windows style pathname, so problems arise when SMLNJ_HOME is set
590    to a Windows style pathname.  _run-sml now converts SMLNJ_HOME
591    to a POSIX pathname on cygwin.
592    
593    ----------------------------------------------------------------------
594    Name: Matthias Blume (blume (at) tti - c (dot) org)
595    Date: 2004/12/16 13:35:00 CST
596    Tag: Release_110_52
597    Description:
598    
599    Last-minute changes incorporated into 110.52.  Release tag moved.
600    
601    The changes:
602       - HashString.hashString' -> HashString.hashSubstring
603       - bug fix in UnivariateStats
604    
605    ----------------------------------------------------------------------
606    Name: Matthias Blume (blume (at) tti - c (dot) org)
607    Date: 2004/12/15 23:40:00 CST
608    Tag: blume-20041215-hashSubstring
609    Description:
610    
611    - HashString.hashString' -> HashString.hashSubstring
612    - corresponding changes in atom.sml
613    - "de-compressed" (aka. un-obfuscated) code for UnivariateStats and
614      added some comments
615    
616    ----------------------------------------------------------------------
617    Name: Matthias Blume (blume (at) tti - c (dot) org)
618    Date: 2004/12/15 15:30:00 CST
619    Tag: (Release_110_52)
620    Description:
621    
622    New working version (110.52).  NEW BOOTFILES!
623    
624    ----------------------------------------------------------------------
625    Name: Matthias Blume (blume (at) tti - c (dot) org)
626    Date: 2004/12/15 12:45:00 CST
627    Tag: blume-20041215-spaces
628    Description:
629    
630    More on the space problem (this time for Win32).
631    
632    ----------------------------------------------------------------------
633    Name: Matthias Blume (blume (at) tti - c (dot) org)
634    Date: 2004/12/14 17:30:00 CST
635    Tag: blume-20041214-spaces
636    Description:
637    
638    Hacked some of the scripts (in particular: the installer) to cope with
639    spaces in filenames a bit better.  But beware: the current "solution"
640    is likely still full of bugs and inherently incomplete.  (We need to
641    do away with those shell scripts for a comprehensive solution.)
642    
643    ----------------------------------------------------------------------
644    Name: Matthias Blume (blume (at) tti - c (dot) org)
645    Date: 2004/12/13 14:45:00 CST
646    Tag: blume-20041213-ml-makedepend
647    Description:
648    
649    Fixed bug in code for ml-makedepend.
650    
651    ----------------------------------------------------------------------
652    Name: Matthias Blume (blume (at) tti - c (dot) org)
653    Date: 2004/12/09 16:30:00 CST
654    Tag: blume-20041209-statistics
655    Description:
656    
657    Added two simple but potentially useful statistics modules to SML/NJ Library.
658    (See CHANGES file there.)
659    
660    ----------------------------------------------------------------------
661    Name: Matthias Blume (blume (at) tti - c (dot) org)
662    Date: 2004/12/01 16:50:00 CST
663    Tag: blume-20041201-atom
664    Description:
665    
666    smlnj-lib:
667    
668      Added function HashString.hashString' for substrings.
669      Hand-inlined CharVector.foldl into HashString (for speed).
670      Modified implementation of structure Atom to avoid extracting
671      strings from substrings unless necessary.
672      (Also see CHANGES file for smlnj-lib.)
673    
674    ----------------------------------------------------------------------
675    Name: Matthias Blume (blume (at) tti - c (dot) org)
676    Date: 2004/11/24 22:15:00 CST
677    Tag: blume-20041124-cml
678    Description:
679    
680    Made sure CML compiles when Position = Int64.
681    
682    ----------------------------------------------------------------------
683    Name: Matthias Blume (blume (at) tti - c (dot) org)
684    Date: 2004/11/24 14:30:00 CST
685    Tag: blume-20041124-position
686    Description:
687    
688    The compiler can now be compiled in a mode that makes structure Position
689    equal to Int64.  The default, however, is unchanged (Position = Int31)
690    for the time being.
691    
692    To enable 64-bit positions, use the following procedure:
693    
694    1. Start sml
695    2. Autoload $smlnj/cmb.cm (if not already autoloaded)
696    3. Type
697          #set (CMB.symval "USE_64_BIT_POSITIONS") (SOME 1);
698    4. Run CMB.make() as usual.
699    
700    This is barely tested.  The only test so far was a little SML program
701    counting the number of characters in an 8-gigabyte file by
702    reading it character-by-character.  That test was successful.
703    
704    In support of 64-bit positions, a number of new functions have been
705    added to the runtime system.
706    
707    ----------------------------------------------------------------------
708    Name: Matthias Blume (blume (at) tti - c (dot) org)
709    Date: 2004/11/23 14:45:00 CST
710    Tag: blume-20041123-useFile
711    Description:
712    
713    Fixed a problem with unhelpful error messages related to problems with
714    .cm- or .sml files that appear as part of the sml command line.
715    
716    ----------------------------------------------------------------------
717    Name: Matthias Blume (blume (at) tti - c (dot) org)
718    Date: 2004/11/18 15:40:00 CST
719    Tag: Release_110_51
720    Description:
721    
722    New working version (110.51).  NEW BOOTFILES!
723    
724    ----------------------------------------------------------------------
725    Name: Matthias Blume (blume (at) tti - c (dot) org)
726    Date: 2004/11/18 15:35:00 CST
727    Tag: <none>
728    Description:
729    
730    Enabled dlopen and friends for FreeBSD (as recommended by Johannes 5
731    Joemann).
732    
733    ----------------------------------------------------------------------
734    Name: Allen Leung (leunga (at) reservoir (dot) com)
735    Date: 2004/11/17 16:05:21 EST 2004
736    Tag: leunga-20041117-mlrisc-live-kill
737    Description:
738    
739       Added support for MLTree constructs LIVE and KILL
740    to all the architectures.
741    
742    ----------------------------------------------------------------------
743    Name: Matthias Blume (blume (at) tti - c (dot) org)
744    Date: 2004/11/13 00:20:00 CST
745    Tag: blume-20041113-versiontool
746    Description:
747    
748    - Stripped down the versiontool: It now only handles the version number.
749      The date string is generated at bootstrap time (during makeml).
750    
751    - In a previous commit, fixed a minor issue with how polyequal is being
752      translated.  In particular, the code now "looks through" abstractions.
753      This results in slightly fewer polyEqual warnings and hopefully slightly
754      more efficient code.  Important examples for where this matters are
755      the new int64 and word64 types.
756    
757    ----------------------------------------------------------------------
758    Name: Matthias Blume (blume (at) tti - c (dot) org)
759    Date: 2004/11/12 00:30:00 CST
760    Tag: blume-20041112-int64
761    Description:
762    
763    Structure Int64 fully hooked in.  (The implementation is not very
764    efficient, though.)
765    
766    ----------------------------------------------------------------------
767    Name: Matthias Blume (blume (at) tti - c (dot) org)
768    Date: 2004/11/11 17:30:00 CST
769    Tag: blume-20041111-more64
770    Description:
771    
772    All the pieces of Word64 are now there, with the exception of the
773    conversions from and to LargeWord. (Eventually these need to be identities,
774    but for the time being they don't even make sense because LargeWord is
775    32-bit wide.)
776    
777    Also started to add similar support for Int64, but major pieces of that
778    are still missing.
779    
780    ----------------------------------------------------------------------
781    Name: Matthias Blume (blume (at) tti - c (dot) org)
782    Date: 2004/11/11 00:15:00 CST
783    Tag: blume-20041111-word64
784    Description:
785    
786    Structure Word64 is now (almost) complete, word literals and patterns
787    seem to work.  There are a few odd pieces missing.  In particular,
788    I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
789    at the moment.
790    
791    Making Word64 official would mean that LargeWord becomes Word64.  But
792    this requires extreme care because most word-word conversions have to
793    go through LargeWord, so making a mistake means loss of efficiency or
794    worse.  Eventually there will be a solution similar to (but actually
795    simpler than) what I did with IntInf.
796    
797    ----------------------------------------------------------------------
798    Name: Matthias Blume (blume (at) tti - c (dot) org)
799    Date: 2004/11/10 18:12:00 CST
800    Tag: blume-20041110-64bit
801    Description:
802    
803    More 64-bit hacking (but still not even half-way there yet).
804    Also, some assorted improvements to the handling of 8-bit words.
805    
806    ----------------------------------------------------------------------
807    Name: Matthias Blume (blume (at) tti - c (dot) org)
808    Date: 2004/11/09 17:50:00 CST
809    Tag: <none>
810    Description:
811    
812    Started adding some infrastructure for supporting 64-bit int- and
813    word-types.  (Still in its very early stages.)
814    
815    ----------------------------------------------------------------------
816    Name: Matthias Blume (blume (at) tti - c (dot) org)
817    Date: 2004/10/28 10:45:00 CDT
818    Tag: Release_110_50
819    Description:
820    
821    New working version (110.50).  NEW BOOTFILES!
822    
823                 =====================
824    
825    Also:
826    
827      - Changend config/srcarchiveurl from a file just
828        containing the URL string into a file containing
829        shell script code.  The code has access to the $VERSION variable.
830      - Made corresponding changes to config/install.sh and config/unpack.
831      - Default contents of config/srcarchiveurl uses $VERSION and
832        normally does not have to be edited to reflect a version change.
833    
834      (As a result, a version change can be done by just editing
835       config/version, the rest is now automatic.)
836    
837    ----------------------------------------------------------------------
838    Name: Matthias Blume (blume (at) tti - c (dot) org)
839    Date: 2004/10/27 17:50:00 CDT
840    Tag: blume-20041027-btrace-msg
841    Description:
842    
843    BackTrace.monitor now also reports the source of the exception that
844    triggered the trace.
845    
846    ----------------------------------------------------------------------
847    Name: Matthias Blume (blume (at) tti - c (dot) org)
848    Date: 2004/10/27 17:20:00 CDT
849    Tag: blume-20041027-x86-c-calls
850    Description:
851    
852    This is the HISTORY entry for two earlier commits, both concerning
853    the x86 c-calls code in MLRISC:
854    
855      - added a missing LOAD in the code that deals with struct arguments
856      - made sure the caller does not add the wrong number of bytes to the
857        stack pointer after a call of a function returning a struct
858        (the callee already pops the implicit argument which points to
859         the space reserved for the result)
860    
861    ----------------------------------------------------------------------
862    Name: Allen Leung (leunga (at) reservoir (dot) com)
863    Date: 2004/10/24 14:00:00 EST
864    Tag: leunga-20041024-x86-gas-fucomip
865    Description:
866    
867       John discovered a bug in the syntax of fucomip.
868       The opcodes FU?COMIP? have been changed to
869    
870           fu?comip? %st(i), %st
871    
872    ----------------------------------------------------------------------
873    Name: Matthias Blume (blume (at) tti - c (dot) org)
874    Date: 2004/10/20 15:06:00 CDT
875    Tag: blume-20041020-standalone-backtrace
876    Description:
877    
878    Added a mechanism for getting back-trace information from standalone
879    programs.  Here is how it works:
880    
881       1. The part of the program from which you want to get backtrace
882          information (usually the whole program) should be wrapped with
883          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
884          main program could be modified from something like
885    
886             fun main (pgm, args) = ...
887    
888         to
889    
890            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
891    
892       2. To be able to access BackTrace.monitor, you have to add
893          library $smlnj-tdp/plugins.cm to the .cm file that contains your
894          main function.
895    
896       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
897          CM might have created in the past for your project).
898    
899       4. Build the system using this command line:
900    
901             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
902                      myprog.cm MyProg.main myprog
903    
904          instead of the usual
905    
906             ml-build myprog.cm MyProg.main myprog
907    
908    I changed a library name:
909    
910       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
911    
912    New libraries:
913    
914       $smlnj-tdp/back-trace.cm
915          -- when loaded causes the back-trace plugin to be installed
916       $smlnj-tdp/coverage.cm
917          -- when loaded causes the coverage plugin to be installed
918    
919    ----------------------------------------------------------------------
920    Name: Matthias Blume (blume (at) tti - c (dot) org)
921    Date: 2004/10/18 16:45:00 CDT
922    Tag: blume-20041018-groupowner
923    Description:
924    
925    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
926    
927    Eliminated group owner specs from .cm files throughout the source tree.
928    
929    ----------------------------------------------------------------------
930    Name: Matthias Blume (blume (at) tti - c (dot) org)
931    Date: 2004/10/15 15:45:00 CDT
932    Tag: blume-20041015-coverage
933    Description:
934    
935    * Test coverage tool added!
936    * Further reorganization of tracing-, debugging-, and profiling support:
937    
938        - moved original BTImp -- now called BackTrace -- into a separate
939          library called $/trace-debug-profile.cm
940        - eliminated all mentions of BTrace from SMLofNJ.Internals
941        - only the instrumentation mechanism is now left in the compiler proper
942        - BackTrace module is a plugin which is NOT plugged in by default
943        - Coverage module is another such plugin
944    
945    To get the benefits of any of these plugin modules, the code in
946    question must be compiled with tdp instrumentation turned on.  This
947    can be done by setting SMLofNJ.Internals.TDP.mode to true.
948    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
949    
950    Plugins are selected at link time.  (Pre-compiled instrumented code
951    can be re-loaded with different plugins in effect.)  When an
952    instrumented module is linked, whatever plugins are at that time
953    enabled will come into effect for that module.
954    
955    To enable the back-trace plugin, load library $/trace-debug-profile.cm
956    and invoke BackTrace.install() (e.g., from the interactive prompt).
957    To enable the coverage plugin, load the same library and invoke
958    Coverage.install().
959    
960    Back-traces are generated automatically on uncaught exceptions and
961    when the code in question explicitly invokes BackTrace.trigger().
962    
963    Coverage (and execution frequency-) information must be queried
964    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
965    
966    ----------------------------------------------------------------------
967    Name: Matthias Blume (blume (at) tti - c (dot) org)
968    Date: 2004/10/14 17:40:00 CDT
969    Tag: blume-20041014-tdp-core
970    Description:
971    
972    Snapshot of a significant overhaul of how the trace/debug/profile support
973    is hooked into the system (specifically: Core and SMLofNJ.Internals).
974    
975    ----------------------------------------------------------------------
976    Name: Matthias Blume (blume (at) tti - c (dot) org)
977    Date: 2004/10/13 16:34:00 CDT
978    Tag: blume-20041013-tdp
979    Description:
980    
981    Some rationalization of names:
982    
983       structure BTrace -> structure TDPInstrument
984       etc.
985    
986    This is is preparation of using the original back-trace
987    instrumentation for other purposes.  "TDP" stands for
988    Trace/Debug/Profile.
989    
990    The control flag controlling whether instrumentation is on or off is now
991    registered under a different name, so instead of running sml as
992    
993     sml -Cinstrument.btrace-mode=true
994    
995    one has to say
996    
997     sml -Ctdp.instrument=true
998    
999    ----------------------------------------------------------------------
1000    Name: Matthias Blume (blume (at) tti - c (dot) org)
1001    Date: 2004/10/11 16:37:00 CDT
1002    Tag: blume-20041011-regions
1003    Description:
1004    
1005    Made some minor modifications to elabcore.sml to have source regions
1006    be propagated more tightly -- resulting in better (i.e., smaller)
1007    regions being reported in error- and debug messages.
1008    
1009    ----------------------------------------------------------------------
1010    Name: Matthias Blume (blume (at) tti - c (dot) org)
1011    Date: 2004/10/08 22:50:00 CDT
1012    Tag: blume-20041008-cmkw
1013    Description:
1014    
1015    Fixed handling of keywords in .cm files: After seeing "is" the lexer
1016    treats subsequent occurrences of "group", "library", "source", "is",
1017    "*", and "-" as ordinary identifiers rather than keywords.
1018    
1019    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
1020    is supposed to accept a tool argument called "source", but this did
1021    not work because of the clash with the keyword.
1022    
1023    ----------------------------------------------------------------------
1024    Name: Matthias Blume (blume (at) tti - c (dot) org)
1025    Date: 2004/10/07 16:00:00 CDT
1026    Tag: blume-20041007-cleanup
1027    Description:
1028    
1029    Assorted cleanup work:
1030    
1031      - got rid of intstrmap in favor of using the library's
1032        hash table implementation
1033      - threw out most of the pathnames stuff, as it was not used anyway
1034      - simplified tokentable implementation
1035      - fixed some minor spelling errors
1036    
1037    ----------------------------------------------------------------------
1038    Name: Matthias Blume (blume (at) tti - c (dot) org)
1039    Date: 2004/10/06 15:15:15 CDT
1040    Tag: blume-20041006-handler
1041    Description:
1042    
1043    Cleaned up the absyn to reflect the invariant that HANDLE always
1044    carries a FNexp as part of the type definition.  This eliminates some
1045    superfluous sanity checks at runtime down the road.
1046    
1047    Some minor cleanup of the btrace code.
1048    
1049    ----------------------------------------------------------------------
1050    Name: Matthias Blume (blume (at) tti - c (dot) org)
1051    Date: 2004/10/01 10:20:30 CDT
1052    Tag: blume-20041001-slave
1053    Description:
1054    
1055    Added hack to make slave mode work in the presence of the version
1056    tool.  (Still, since the master does two passes over the code for
1057    CMB.make, the release number gets bumped twice when slaves are
1058    attached.  I don't know if this is worth fixing...)
1059    
1060    ----------------------------------------------------------------------
1061    Name: Matthias Blume (blume (at) tti - c (dot) org)
1062    Date: 2004/09/30 10:55:00 CDT
1063    Tag: blume-20040930-version
1064    Description:
1065    
1066    * Moved the "version" magic into its own little library under
1067      src/system/smlnj/internal.  This avoids expensive reconstruction of
1068      a stable src/compiler/core.cm.
1069    
1070    * At the same time, structure CompilerVersion is now known as
1071      structure SMLNJVersion.
1072    
1073    * Arranged for the version tool to NOT kick in when rebuilding the system
1074      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
1075      Also, loading the versiontool does not work when rebuilding the system
1076      because CM is not properly initialized at that time.
1077    
1078    ----------------------------------------------------------------------
1079    Name: Matthias Blume (blume (at) tti - c (dot) org)
1080    Date: 2004/09/29 14:00:00 CDT
1081    Tag: blume-20040929-autoversion
1082    Description:
1083    
1084    Implemented some CM magic to have
1085        file src/compiler/TopLevel/main/version.sml
1086    generated automagically.
1087    The version is taken from two files: config/version and config/release.
1088    The first is expected to contain a two-part version number such as 110.49.
1089    The second should contain a single number, but it may be missing.
1090    
1091    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
1092    time the version tool is loaded (which is the first time you say CMB.make),
1093    then the tool will increment the value stored in config/release every
1094    time CMB.make is invoked.
1095    
1096    The binfile format is now insensitive to anything beyond the first
1097    two components of a version number, so bumping the release does not render
1098    binfiles incompatible.  Auto-bumping can be used to keep track of versions
1099    during development without invalidating existing binfiles.
1100    
1101    In any case, every CMB.make updates the date information in version.sml.
1102    (This is the date that is printed in the banner.)
1103    
1104    ----------------------------------------------------------------------
1105    Name: Matthias Blume (blume (at) tti - c (dot) org)
1106    Date: 2004/09/28 10:53:00 CDT
1107    Tag: blume-20040928-controls
1108    Description:
1109    
1110    Some cleanup of the controls code.
1111    
1112    ----------------------------------------------------------------------
1113    Name: Matthias Blume (blume (at) tti - c (dot) org)
1114    Date: 2004/09/27 22:08:00 CDT
1115    Tag: blume-20040927-controls
1116    Description:
1117    
1118    Added two pieces of functionality to the Controls interface:
1119    
1120    1.   val save'restore: 'a control -> unit -> unit
1121    
1122       grabs the current value of the control in stage 1 and restores it
1123       in stage 2.
1124    
1125    2.   val set' : 'a control * 'a -> unit -> unit
1126    
1127       stores the given value into the control in stage 2 (i.e., delayed)
1128       but does all error checking in stage 1.
1129       (This is for string controls that need to do parse their argument --
1130        something that might fail.  In some cases, notably in CM, one
1131        already knows the intended argument but wants to delay the actual
1132        assignment until a time when error recovery would be more difficult.)
1133    
1134    Changed the handling of controls in tool arguments to classes "sml" and
1135    "lazysml":
1136       - use Controls.save'restore as a more robust way of restoring the
1137         old value (in particular: without having to re-parse the string)
1138       - use controls to handle the "overload" keyword in the init group
1139         (I believe this change actually fixes a long-standing obscure bug.)
1140    
1141    ----------------------------------------------------------------------
1142    Name: Matthias Blume (blume (at) tti - c (dot) org)
1143    Date: 2004/09/27 17:00:00 CDT
1144    Tag: blume-20040927-lazysml
1145    Description:
1146    
1147    Added a new tool class called "lazysml" to CM's tool chest.  The only
1148    difference to "sml" is that compilation is done with Control.lazysml
1149    set to true.  A source of class "lazysml" is automatically recognized
1150    by a file name suffix of ".lml".
1151    
1152    In addition to the above feature, the original class "sml" now also
1153    supports a tool argument "lazy" which has the same effect.  As a
1154    result, the following three lines are equivalent:
1155    
1156        foo.sml : lazysml
1157        foo.sml : sml (lazy)
1158        foo.sml (lazy)
1159    
1160    The setting goes into effect both during parsing and during
1161    compilation.  The original setting is restored right after parsing and
1162    after compilation, respectively.
1163    
1164    In addition to all the above, there is also a general mechanism to set
1165    ANY of the "controls" that are available at the command line via
1166    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
1167    apply as well. (In fact, "lazy" is implemented as a special case of
1168    the general mechanism.)
1169    
1170    The .cm file syntax uses a new keyword tool argument called "with".
1171    There are several ways of indicating the desired settings:
1172    
1173        foo.sml (with:parser.quotations=true)
1174        foo.sml (with:(name:parser.quotations value:true))
1175        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
1176        foo.sml (with:(name1=value1 name2=value2 ...))
1177        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
1178    
1179        etc.
1180    
1181    Another possible abbreviation is to leave out the =v or value:v part
1182    if the name refers to a boolean control (in which case the value is
1183    taken to be true).  Thus, one could get lazy sml also by saying:
1184    
1185        foo.sml (with:parser.lazy-keyword=true)
1186        foo.sml (with:parser.lazy-keyword)
1187        foo.sml (with:(name:parser.lazy-keyword value:true))
1188        foo.sml (with:(name:parser.lazy-keyword))
1189    
1190    ----------------------------------------------------------------------
1191    Name: Matthias Blume (blume (at) tti - c (dot) org)
1192    Date: 2004/09/24 16:22:00 CDT
1193    Tag: blume-20040924-ppc-long-branch
1194    Description:
1195    
1196    Turned message about "emiting long form of branch" off by default.
1197    Added a control flag to turn it back on when desired.
1198    
1199    ----------------------------------------------------------------------
1200    Name: Matthias Blume (blume (at) tti - c (dot) org)
1201    Date: 2004/09/24 16:05:00 CDT
1202    Tag: blume-20040924-rounding
1203    Description:
1204    
1205    Applied patch for setting rounding modes under Mac OS X.  Thanks to
1206    Melissa O'Neill for providing the code!
1207    
1208    ----------------------------------------------------------------------
1209    Name: Matthias Blume (blume (at) tti - c (dot) org)
1210    Date: 2004/09/23 17:30:00 CDT
1211    Tag: blume-20040923-envvars
1212    Description:
1213    
1214    1. Changed definition of type ControlRegistry.registry_tree to
1215       include control_info (i.e., the name of the controlling environment
1216       variable).
1217    
1218    2. Added command-line flags -e and -E to print the names of environment
1219       variables that can be used to control internal settings.
1220       (This uses the new API mentioned in 1.)
1221    
1222    ----------------------------------------------------------------------
1223    Name: Matthias Blume (blume (at) tti - c (dot) org)
1224    Date: 2004/09/13 16:50:00 CDT
1225    Tag: Release_110_49
1226    Description:
1227    
1228    New working version (110.49).  NEW BOOTFILES!
1229    
1230    ----------------------------------------------------------------------
1231    Name: Matthias Blume
1232    Date: 2004/09/13 16:20:00 CDT
1233    Tag: blume-20040913-config-mlrisc
1234    Description:
1235    
1236    Put target "mlrisc" back into the default list.
1237    (There is no harm in having it, and some users have expressed their
1238    wish to have "mlrisc" included by default.)
1239    
1240    ----------------------------------------------------------------------
1241    Name: John Reppy
1242    Date: 2004/09/13
1243    Tag: jhr-20040913-signals
1244    Description:
1245    
1246    Fixed the signal masking code to properly nest mask/unmask operations
1247    on a per-signal basis.
1248    
1249    ----------------------------------------------------------------------
1250    Name: Matthias Blume (blume (at) tti - c (dot) org)
1251    Date: 2004/09/08 13:20:00 CDT
1252    Tag: blume-20040908-heap-magic
1253    Description:
1254    
1255    Bumped the heap macig number to 0x09082004 to account for the changed
1256    layout of the ML frame under MacOS X.
1257    
1258    ----------------------------------------------------------------------
1259    Name: Allen Leung (leunga (at) reservoir (dot) com)
1260    Date: 2004/09/03 11:26:00 EST
1261    Tag: leunga-20040903-cygwin-install
1262    Description:
1263    
1264       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
1265    The Cygwin runtime is turned on by setting the environment
1266    variable SMLNJ_CYGWIN_RUNTIME to 1.
1267    
1268    ----------------------------------------------------------------------
1269    Name: Matthias Blume (blume (at) tti - c (dot) org)
1270    Date: 2004/08/31 17:15:00 CDT
1271    Tag: blume-20040831-core
1272    Description:
1273    
1274    Added some exports to src/compiler/core.cm upon request by J. Joemann.
1275    
1276    ----------------------------------------------------------------------
1277    Name: Matthias Blume (blume (at) tti - c (dot) org)
1278    Date: 2004/08/30 17:55:00 CDT
1279    Tag: blume-20040830-installer
1280    Description:
1281    
1282    Upon request by Johannes Joemann:
1283    
1284    - improved ML code of installer to fall back to coping when renaming
1285      fails (i.e., when source and target are on different file systems);
1286      the code compiles but has yet to be tested in anger
1287    - removed mlrisc from list of default targets (config/targets)
1288    
1289    ----------------------------------------------------------------------
1290    Name: Matthias Blume (blume (at) tti - c (dot) org)
1291    Date: 2004/08/27 17:20:00 CDT
1292    Tag: blume-20040827-ptreql
1293    Description:
1294    
1295    Added ptreql primop to structure InlineT (upon request from Larry
1296    Paulson).
1297    
1298    ----------------------------------------------------------------------
1299    Name: Allen Leung (leunga (at) reservoir (dot) com)
1300    Date: 2004/08/15 21:21:00 EST
1301    Tag: leunga-110_48-udgraph
1302    Description:
1303    
1304    Another bug fix from Carl Hauser:
1305    
1306    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
1307    > 48c48
1308    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
1309    > e::L)
1310    > ---
1311    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
1312    > Without this, any deletion of an edge in an undirected graph does severe
1313    > violence to the graph.
1314    
1315    ----------------------------------------------------------------------
1316    Name: Allen Leung (leunga (at) reservoir (dot) com)
1317    Date: 2004/08/10 23:35:00 EST
1318    Tag: leunga-110_48-ppc
1319    Description:
1320    
1321       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
1322    
1323    ----------------------------------------------------------------------
1324    Name: Matthias Blume (blume (at) tti - c (dot) org)
1325    Date: 2004/08/10 12:00:00 CDT
1326    Tag: Release_110_48
1327    Description:
1328    
1329    New working version (110.48).  NEW BOOTFILES!
1330    
1331    ----------------------------------------------------------------------
1332    Name: Allen Leung (leunga (at) reservoir (dot) com)
1333    Date: 2004/08/09 12:21:00 EST
1334    Tag: leunga-110_47-dijsktra
1335    Description:
1336    
1337    Bug fix from Carl Hauser:
1338    
1339    single_source_shortest_paths in dijkstra.sml was observed to get wrong
1340    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
1341    
1342    The problem is that following the expression A.update(dist,s,Num.zero)
1343    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
1344    
1345    ----------------------------------------------------------------------
1346    Name: Matthias Blume (blume (at) tti - c (dot) org)
1347    Date: 2004/08/06 18:10:00 CDT
1348    Tag: blume-20040806-cmdline
1349    Description:
1350    
1351    Fiddled with handling of command-line options:
1352    
1353      * sml now quits after processing the command line
1354        if -H, -S, -h<n>, or -s<n> appears as the last
1355        command-line argument
1356      * a new option -q terminates the session when encountered on
1357        the command line; subsequent arguments will be ignored
1358      * bug fixes: short (erroneous) arguments are no longer ignored
1359        completely
1360    
1361    ----------------------------------------------------------------------
1362    Name: Allen Leung (leunga (at) reservoir (dot) com)
1363    Date: 2004/08/04 18:17:00 EST
1364    Tag: leunga-110_47-ppc-ibm-asm
1365    Description:
1366    
1367      - Added minimal IBM assembly syntax support for PowerPC.
1368    
1369      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
1370        been altered in the runtime.  We need an automatic way to keep the file
1371        in sync.
1372    
1373    ----------------------------------------------------------------------
1374    Name: Matthias Blume (blume (at) tti - c (dot) org)
1375    Date: 2004/08/04 14:00:00 CDT
1376    Tag: Release_110_47
1377    Description:
1378    
1379    New working version (110.47).  NEW BOOTFILES!
1380    
1381    ----------------------------------------------------------------------
1382    Name: Matthias Blume (blume (at) tti - c (dot) org)
1383    Date: 2004/08/03 14:25:00 CDT
1384    Tag: blume-20040803-callingconv
1385    Description:
1386    
1387    Added low-level support for choosing C calling conventions by
1388    twiddling the type of rawccall.  (See
1389    src/compiler/Semant/types/cproto.sml for details.)
1390    
1391    ----------------------------------------------------------------------
1392    Name: Matthias Blume (blume (at) tti - c (dot) org)
1393    Date: 2004/08/02 15:55:00 CDT
1394    Tag: blume-20040802-backout
1395    Description:
1396    
1397    Backed out of change to win32-filesys.c.  The earlier patch to
1398    get_file_time caused CM to produce files with the wrong time stamp.
1399    
1400    ----------------------------------------------------------------------
1401    Name: Matthias Blume (blume (at) tti - c (dot) org)
1402    Date: 2004/08/02 14:45:00 CDT
1403    Tag: blume-20040802-nlffi-win32
1404    Description:
1405    
1406    Added NLFFI support for Win32, adapted from a patch provided by David
1407    Hansel.  This is currently completely untested.  Also, the issue
1408    concerning stdcall vs. ccall is still unresolved.
1409    
1410    ----------------------------------------------------------------------
1411    Name: Matthias Blume (blume (at) tti - c (dot) org)
1412    Date: 2004/07/30 17:55:00 CDT
1413    Tag: blume-20040730-various
1414    Description:
1415    
1416    Gearing up towards 110.47...
1417    
1418    - various minor bugfixes to ml-nlffigen
1419    - a beginning of a manual for nlffi
1420    
1421    - eliminated 'export name=value' in config/install.sh as this does
1422      not work with certain versions of /bin/sh
1423      (Thanks to David King at Motorola for catching this.)
1424    
1425    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
1426      - added a test for tm==NULL to gmtime.c and localtime.c
1427      - applied patch for incorrect GetFileTime under win32
1428      - toSeconds -> toMilliseconds in Win32/win32-process.sml
1429    
1430    ----------------------------------------------------------------------
1431    Name: Matthias Blume (blume (at) tti - c (dot) org)
1432    Date: 2004/07/21 18:20:00 CDT
1433    Tag: blume-20040721-nlffigen
1434    Description:
1435    
1436    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
1437      for a typedef to an incomplete type, but leave out the "typ" member.
1438      (This is just for consistency.)
1439    - Started to produce what is supposed to become better (i.e., comprehensive)
1440      documentation of what ml-nlffigen does and produces.
1441    
1442    ----------------------------------------------------------------------
1443    Name: Matthias Blume (blume (at) tti - c (dot) org)
1444    Date: 2004/07/14 16:25:00 CDT
1445    Tag: blume-20040714-union
1446    Description:
1447    
1448    Added C_UNION to c-calls/c-types.sml and updated the machinery
1449    (ml-nlffigen, cproto.sml) that conveys C function interface
1450    information to the code generator.
1451    
1452    However, the actual architecture-specific implementation of function
1453    arguments and results that are C unions is still not implemented.
1454    
1455    ----------------------------------------------------------------------
1456    Name: Allen Leung (leunga (at) reservoir (dot) com)
1457    Date: 2004/07/14 14:38:00 EST
1458    Tag: leunga-110_46_1-ppc-lwzu
1459    Description:
1460    
1461        Added these instructions to the PowerPC architecture:
1462        LBZU(X), LHZU(X), LWZU(X),
1463        STWU(X), STFDU, STFSU
1464    
1465        etc...
1466    
1467        Note: I haven't added their instruction encoding into the description.
1468    
1469    ----------------------------------------------------------------------
1470    Name: Allen Leung (leunga (at) reservoir (dot) com)
1471    Date: 2004/07/13 15:04:00 EST
1472    Tag: leunga-110_46_1-ppc-lwarx
1473    Description:
1474    
1475       Added the two instructions LWARX and STWCX to the PowerPC
1476    instruction set.
1477    
1478       A (untested) rewrite of loop-structure.sml.  The old version
1479    is completely broken.
1480    
1481    ----------------------------------------------------------------------
1482    Name: Matthias Blume (blume (at) tti - c (dot) org)
1483    Date: 2004/07/13 13:50:00 CDT
1484    Tag: blume-20040713-nlffi
1485    Description:
1486    
1487    - use paramAlloc to report c-calls with too many arguments
1488      (for PPC version where parameter area is pre-allocated)
1489    - added ccall_maxargspace to machspec (to implement the above)
1490    - made "make" commend in CM's "make" tool configurable
1491    - added option (default: on) for passing the name of the SML/NJ's "bin"
1492      directory to "make"; the call looks like this:
1493    
1494         make <options> SMLNJ_BINDIR=<dir> <target>
1495    
1496      This can be used by the Makefile to, e.g., pick the "right" version
1497      of ml-nlffigen.
1498    - minor code tweaks
1499    
1500    ----------------------------------------------------------------------
1501    Name: Matthias Blume (blume (at) tti - c (dot) org)
1502    Date: 2004/07/12 22:50:00 CDT
1503    Tag: blume-110_46_1-macosx-nlffi
1504    Description:
1505    
1506    NLFFI under Mac OS X now working (sort of).  This is largely untested,
1507    though.
1508    
1509    Note:  1. You have to make a new, clean build of the runtime system.
1510           2. There are new BOOTFILES, you have to use them!
1511              (Doing the bootstrap process yourself would be *very* painful!
1512               If you absolutely have to do it, build the system under
1513               a different architecture and then cross-compile.)
1514    
1515    Version bumped to 110.46.1 to account for runtime data format changes.
1516    
1517    ----------------------------------------------------------------------
1518    Name: Matthias Blume (blume (at) tti - c (dot) org)
1519    Date: 2004/06/18 14:30:00 CDT
1520    Tag: blume-20040618-unix
1521    Description:
1522    
1523    Changed the implementation of structure Unix so that the same stream
1524    is returned every time one of the {text,bin}{In,Out}streamOf functions
1525    is invoked on the same proc.  This is not what the spec currently
1526    says -- although IMO it arguably should.  (See discussion below.)
1527    
1528    ----------------------------------------------------------------------
1529    Name: Matthias Blume (blume (at) tti - c (dot) org)
1530    Date: 2004/06/17 18:15:00 CDT
1531    Tag: Release_110_46
1532    Description:
1533    
1534    New working version (110.46).  NEW BOOTFILES!
1535    
1536    ----------------------------------------------------------------------
1537    Name: Matthias Blume (blume (at) tti - c (dot) org)
1538    Date: 2004/06/17 17:20:00 CDT
1539    Tag: blume-20040617-timer-unix
1540    Description:
1541    
1542    Changed the interface of structures Timer and Unix to match the most
1543    recent Basis spec.
1544    
1545    In the case of Unix there still seems to be an open/weird issue:
1546    
1547       The {text,bin}{In,Out}streamOf functions are supposed to create
1548       fresh streams whenever they are called -- as opposed to have them
1549       return the same stream every time.  This design is supposed to
1550       prevent space leaks caused by proc values hanging on to streams.
1551    
1552       The reap function, on the other hand, is supposed to close the
1553       streams.  This cannot be done without having a handle on the
1554       stream in proc after all...
1555    
1556    I took the liberty to implement the following stopgap solution:
1557    
1558      The proc value hangs on to the most recently created stream(s).
1559      Reap closes those.  If either or both of the two streams hadn't
1560      been created at all yet, then reap will close the corresponding
1561      file descriptors directly.
1562    
1563    PS: I don't understand the original space leak argument anymore.  If
1564    a proc hangs on to the imperative stream, then I/O operations on those
1565    will advance the state of the cached stream and avoid the space leak.
1566    
1567    ----------------------------------------------------------------------
1568    Name: Matthias Blume (blume (at) tti - c (dot) org)
1569    Date: 2004/05/28 16:45:00 CDT
1570    Tag: blume-20040528-basis
1571    Description:
1572    
1573    Added signature PACK_REAL and exported functor PrimIO.
1574    
1575    ----------------------------------------------------------------------
1576    Name: Matthias Blume (blume (at) tti - c (dot) org)
1577    Date: 2004/05/25 16:00:00 CDT
1578    Tag: blume-20040525-group-owner
1579    Description:
1580    
1581    CM now ignores (but still accepts) the "owner" information in group
1582    descriptions.  The owner of a group is its next enclosing
1583    library. Each group must have a unique owner.  (There is a virtual
1584    "toplevel" library that owns groups which are not nested within a real
1585    library.)  Previously, each group had to explicitly declare its owner,
1586    and CM would check that such a declaration is correct.  The new scheme
1587    is to have CM check that for each group there is precisely one owning
1588    library.
1589    
1590    The advantage of the new scheme is that the programmer no longer needs
1591    to maintain the somewhat annoying owner information.  The downside is
1592    that CM cannot enforce the ownership rule across multiple runs of
1593    CM.make.  Fortunately, enclosing the same group in two different
1594    libraries A and B which are not part of the same program does not
1595    cause real problems.
1596    
1597    ----------------------------------------------------------------------
1598    Name: Matthias Blume (blume (at) tti - c (dot) org)
1599    Date: 2004/05/20 16:00:00 CDT
1600    Tag: blume-20040520-win32
1601    Description:
1602    
1603    Made the win32 version work again.  (Strangely, a misplaced comma had
1604    slipped into win32-process.c which prevented the runtime from being
1605    compiled correctly.)
1606    
1607    Also, included a minor addition to ml-build.bat analogous to what was
1608    done in blume-20040519-ml-build.
1609    
1610    ----------------------------------------------------------------------
1611    Name: Matthias Blume (blume (at) tti - c (dot) org)
1612    Date: 2004/05/19 22:10:00 CDT
1613    Tag: blume-20040519-ml-build
1614    Description:
1615    
1616    Arranged for ml-build to clean up after itself a little bit better.
1617    The script generates a temporary SML source file and compiles it using
1618    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
1619    gets rid of those at the end, so they don't accumulate under .cm.
1620    
1621    This required a minor change to install.sh because the name of the
1622    metadata directory (default: .cm) is actually configurable at
1623    installation time.
1624    
1625    ----------------------------------------------------------------------
1626    Name: Matthias Blume (blume (at) tti - c (dot) org)
1627    Date: 2004/05/18 15:50:00 CDT
1628    Tag: blume-20040518-mkreader
1629    Description:
1630    
1631    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
1632    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
1633    
1634    ----------------------------------------------------------------------
1635    Name: Matthias Blume (blume (at) tti - c (dot) org)
1636    Date: 2004/05/11 14:35:00 CDT
1637    Tag: blume-20040511-win32sock
1638    Description:
1639    
1640    Added previously missing support for many socket-related functions
1641    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
1642    for the voluminous patch!
1643    
1644    (I have not tested this patch under win32 yet.)
1645    
1646    Here is David's e-mail:
1647    
1648    Hi,
1649    
1650    Attached to this email you find a diff against sml/nj 110.45
1651    that will enable socket support under Windows.
1652    
1653    To apply the patch (using unix or cygwin)
1654    1) gunzip runtime.diff.gz
1655    2) "cd" into "src/runtime" in the source tree of a fresh
1656       110.45 installation.
1657    3) patch -p 1 < [your/path/to]runtime.diff
1658    
1659    The code compiles fine but has NOT yet been extensively tested.
1660    I only ran a few tests for basic socket client functionality
1661    (which worked fine).  Especially the functions that use ioctl
1662    are not tested at all and might not work (see below).
1663    
1664    I implemented this since we want to move to a newer version of sml/nj
1665    but need socket support in order to use it.  This is the first time I
1666    even had a look at the sml/nj source,  so please review my changes
1667    before making this part of the distribution!  Here are a few issues
1668    that I think might be better for someone to solve who is more
1669    familiar with the sml/nj source (and socket programming):
1670    
1671    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
1672      exception since I could not figure out what the windows equivalent
1673      of these functions is
1674    
1675    - In sockets-osdep.h there are a some #include statements that are
1676      only used in a few files that include sockets-osdep.h
1677    
1678    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
1679      does not process its return value since I don't know how to
1680      report an error upwards.
1681    
1682    - It would probably be good to have a call to WSACleanup() when
1683      the library is unloaded (if there is such a possibility).
1684      Otherwise I think Windows will take care of this automatically
1685      when the process finishes.
1686    
1687    - I used ioctlsocket() as a replacement for ioctl() but I have
1688      no idea if that is actually the proper replacement on Windows.
1689    
1690    - All these issues are marked in the code by "FIXME" comments.
1691    
1692    We use sml/nj extensively in our products and are quite happy
1693    with it.  I hope this contribution will help you.
1694    
1695    Keep up the good work!
1696    
1697    David
1698    
1699    ----------------------------------------------------------------------
1700    Name: Matthias Blume (blume (at) tti - c (dot) org)
1701    Date: 2004/05/11 14:20:00 CDT
1702    Tag: blume-20040511-installml
1703    Description:
1704    
1705    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
1706    the patch.)
1707    
1708    ----------------------------------------------------------------------
1709    Name: Matthias Blume (blume (at) tti - c (dot) org)
1710    Date: 2004/05/11 14:05:00 CDT
1711    Tag: blume-20040511-nlffi-netbsd
1712    Description:
1713    
1714    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
1715    the patch.)
1716    
1717    ----------------------------------------------------------------------
1718    Name: Matthias Blume (blume (at) tti - c (dot) org)
1719    Date: 2004/05/11 12:05:00 CDT
1720    Tag: blume-20040511-exports
1721    Description:
1722    
1723    As per request by Adam Chlipala <adam@hcoop.net>, extended various
1724    export lists in compiler-related .cm-files.
1725    
1726    ----------------------------------------------------------------------
1727    Name: Matthias Blume (blume (at) tti - c (dot) org)
1728    Date: 2004/05/11 11:35:00 CDT
1729    Tag: blume-20040511-allsource
1730    Description:
1731    
1732    The installer now honors the "src-smlnj" target again, although its meaning
1733    has changed from "all sources required for the compiler" to "all sources
1734    the installer knows about".  In other words, if you enable "src-smlnj"
1735    in the "targets" file, then the installer will pull in sources for
1736    everything.  (Notice that this refers to source code only.  Compiled
1737    code is still only installed for modules that were requested explicitly
1738    or which are required for other modules that were requested explicitly.)
1739    
1740    ----------------------------------------------------------------------
1741    Name: Matthias Blume (blume (at) tti - c (dot) org)
1742    Date: 2004/04/23 17:40:00 CDT
1743    Tag: blume-20040423-ieee-scan
1744    Description:
1745    
1746    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
1747    in the exponent calculation we get INF or ZERO (depending on the mantissa
1748    and the sign of the exponent).
1749    
1750    ----------------------------------------------------------------------
1751    Name: Matthias Blume (blume (at) tti - c (dot) org)
1752    Date: 2004/04/23 10:40:00 CDT
1753    Tag: blume-20040423-ml-build
1754    Description:
1755    
1756    The ml-build script now terminates with a non-0 status when something
1757    goes wrong.
1758    
1759    ----------------------------------------------------------------------
1760    Name: Matthias Blume (blume (at) tti - c (dot) org)
1761    Date: 2004/04/22 16:35:00 CDT
1762    Tag: blume-20040422-Option
1763    Description:
1764    
1765    Made exception Option to be the same as exception Option.Option
1766    (as it should be).
1767    
1768    ----------------------------------------------------------------------
1769    Name: Allen Leung (leunga (at) reservoir (dot) com)
1770    Date: 2004/03/19 14:40:00 EST
1771    Tag: leunga-20040319-cygwin-nlffi
1772    Description:
1773    
1774    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1775    of SML/NJ.  The problem is that
1776    
1777       lib = dlopen(NULL, ...)
1778       f   = dlsym(lib, "malloc");
1779    
1780    does not work on Windows unless we explicitly export symbols
1781    such as 'malloc' during linking.  We fixed this by explicitly
1782    exporting the required symbols with the magic gcc incantation:
1783    
1784        -Wl,--export-all cygwin.def
1785    
1786    where cygwin.def is a file containing all the symbols that
1787    we wish to export.
1788    
1789    I suspect this is a Windows problem and we'll have to
1790    do the same (somehow with windows compilers) when
1791    we build the native win32 version with the system
1792    calls LoadLibrary/GetProcAddress.
1793    
1794    ----------------------------------------------------------------------
1795    Name: Matthias Blume (blume (at) tti - c (dot) org)
1796    Date: 2004/03/04 16:35:00 CST
1797    Tag: blume-20040304-intinf-fmt
1798    Description:
1799    
1800    Fixed problem with IntInf.fmt (sign would show up on the right instead
1801    of on the left for BIN, OCT, and HEX).
1802    
1803    ----------------------------------------------------------------------
1804    Name: Matthias Blume (blume (at) tti - c (dot) org)
1805    Date: 2004/03/04 11:25:00 CST
1806    Tag: blume-20040304-symlinks
1807    Description:
1808    
1809    Fixed problem with installer script (unix only) where bin/ml-yacc and
1810    friends pointed (via symlinks) to absolute locations instead of just
1811    .run-sml.  This was reported by Vesa A Norrman.
1812    
1813    ----------------------------------------------------------------------
1814    Name: Matthias Blume (blume (at) tti - c (dot) org)
1815    Date: 2004/02/13 14:50:00 CST
1816    Tag: Release_110_45
1817    Description:
1818    
1819    New working version (110.45). New bootfiles.
1820    
1821    ----------------------------------------------------------------------
1822    Name: Matthias Blume (blume (at) tti - c (dot) org)
1823    Date: 2004/01/26 15:15:15 CST
1824    Tag: blume-20040126-toplevel
1825    Description:
1826    
1827    Improved handling of exceptions at the interactive toplevel.
1828    
1829    ----------------------------------------------------------------------
1830    Name: Matthias Blume (blume (at) tti - c (dot) org)
1831    Date: 2004/01/26 11:25:00
1832    Tag: blume-20040126-app
1833    Description:
1834    
1835    Type of top-level "app" corrected.
1836    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1837    
1838    ----------------------------------------------------------------------
1839    Name: Matthias Blume (blume (at) tti - c (dot) org)
1840    Date: 2003/11/18 17:10 CST
1841    Tag: blume-20031118-basis-fiddle
1842    Description:
1843    
1844    - changed Timer interface to what might become the spec
1845    - POSIX_FLAGS -> BIT_FLAGS according to spec
1846    - some other minor discrepancies wrt. spec eliminated
1847    
1848    ----------------------------------------------------------------------
1849    Name: Matthias Blume (blume (at) tti - c (dot) org)
1850    Date: 2003/11/06 12:00:00 CST
1851    Tag: Release_110_44
1852    Description:
1853    
1854    New working version (110.44). New bootfiles.
1855    
1856    ----------------------------------------------------------------------
1857    Name: Matthias Blume (blume (at) tti - c (dot) org)
1858    Date: 2003/11/04 11:50:00 CST
1859    Tag: blume-20031104-move-libraries
1860    Description:
1861    
1862    Eliminated the "dont_move_libraries" directive in config/targets.
1863    (The mechanism was broken and could not be fixed easily.  Moreover,
1864    there does not seem to be any reason not to move all libraries into
1865    lib during installation.  I originally implemented this directive as a
1866    backward-compatibility feature when I first introduced the new CM.
1867    Now that things have been stable for a long time and going back to the
1868    old CM is not an option, there is no reason to keep it around.)
1869    
1870    ----------------------------------------------------------------------
1871    Name: Matthias Blume (blume (at) tti - c (dot) org)
1872    Date: 2003/11/03 16:00:00 CST
1873    Tag: blume-20031103-installdir
1874    Description:
1875    
1876    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1877    Richards for pointing out the problem and providing the solution.)
1878    
1879    ----------------------------------------------------------------------
1880    Name: Matthias Blume (blume (at) tti - c (dot) org)
1881    Date: 2003/10/01 17:05:00 CDT
1882    Tag: blume-20031001-lal-mlrisc
1883    Description:
1884    
1885    MLRISC bug fix from Lal.
1886    
1887    ----------------------------------------------------------------------
1888    Name: Matthias Blume (blume (at) tti - c (dot) org)
1889    Date: 2003/09/30 16:10:00 CDT
1890    Tag: blume-20030930-primio-bat
1891    Description:
1892    
1893    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1894    2. Improved .bat files (for Win32 port) to make things work under Win95.
1895       (thanks to Aaron S. Hawley for this one)
1896    
1897    ----------------------------------------------------------------------
1898    Name: Matthias Blume (blume (at) tti - c (dot) org)
1899    Date: 2003/09/26 16:05:00 CDT
1900    Tag: blume-20030926-wrappriv
1901    Description:
1902    
1903    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1904    
1905    ----------------------------------------------------------------------
1906    Name: Matthias Blume (blume (at) tti - c (dot) org)
1907    Date: 2003/09/26 15:00:00 CDT
1908    Tag: blume-20030926-110_43_3
1909    Description:
1910    
1911    - additional cleanup
1912    - version number bump, NEW BOOTFILES
1913    
1914    ----------------------------------------------------------------------
1915    Name: Matthias Blume (blume (at) tti - c (dot) org)
1916    Date: 2003/09/26 12:00:00 CDT
1917    Tag: blume-20030926-ppautoload
1918    Description:
1919    
1920    I modified the read-eval-print loop so that the autoloader gets
1921    invoked whenever the prettyprinter tries to look up a symbol that
1922    is not currently defined in the toplevel environment but which
1923    appears in CM's autoload registry.  As a result, we see far fewer of
1924    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1925    
1926    In addition to this I tried to clean up some pieces of the Basis
1927    implementation (e.g., Socket, Word8Array) in order to prevent other
1928    instances of these ?.Foo.Bar.xxx names from being printed.
1929    
1930    The mechanism that picks names for types still needs some work, though.
1931    (Right now it seems that if there is a type A.t which is defined to
1932    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1933    "?.B.u" although the perhaps more sensible solution would be to use
1934    "A.t" in this case.  In other words, the prettyprinter should follow
1935    a chain of DEFtycs not farther than there are corresponding toplevel
1936    names in the current environment.)
1937    
1938    ----------------------------------------------------------------------
1939    Name: Matthias Blume (blume (at) tti - c (dot) org)
1940    Date: 2003/09/24 16:31:00 CDT
1941    Tag: blume-20030924-installer
1942    Description:
1943    
1944    Another installer tweak:  All the ML code for the installer is now
1945    compiled during CMB.make and put into a little library called
1946    $smlnj/installer.cm.  The installation then simply invokes
1947    
1948       sml -m $smlnj/installer.cm
1949    
1950    and everything happens automagically.
1951    
1952    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1953    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1954            and BINDIR.
1955    
1956    The new scheme guarantees that the ML code responsible for the installation
1957    is in sync with the APIs of the main system.  Also, the installer is
1958    somewhat faster because the installer script is precompiled.
1959    
1960    ----------------------------------------------------------------------
1961    Name: Matthias Blume (blume (at) tti - c (dot) org)
1962    Date: 2003/09/24 15:35:00 CDT
1963    Tag: blume-20030924-synsock
1964    Description:
1965    
1966    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1967    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1968    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1969    implementations of non-blocking operations from CML's Socket
1970    structure.
1971    
1972    ----------------------------------------------------------------------
1973    Name: Matthias Blume (blume (at) tti - c (dot) org)
1974    Date: 2003/09/24 15:10:05 CDT
1975    Tag: blume-20030924-sockets
1976    Description:
1977    
1978    1. Fixed SOCKET API and implementation to match Basis spec.
1979       This required changing the internal representation of sockets to one
1980       that remembers (for each socket file descriptor) whether it is currently
1981       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1982       call is made only if the state actually needs to change).
1983    
1984    2. OS-specific details of sockets were moved into separate files, thus
1985       making it possible to unify the bulk of the socket implementations
1986       between Unix and Win32.
1987    
1988    3. CML's socket API changed accordingly.
1989       (Note that we need to remove non-blocking functions from this API
1990       since they are redundant in the case of CML!)
1991    
1992    4. CML's socket implementation now makes use of non-blocking functions
1993       provided by Basis, thus removing all OS-dependent code from this part
1994       of CML.
1995    
1996    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1997    
1998    ----------------------------------------------------------------------
1999    Name: Matthias Blume (blume (at) tti - c (dot) org)
2000    Date: 2003/09/22 12:10:00 CDT
2001    Tag: blume-20030922-110_43_2
2002    Description:
2003    
2004    Made a new interim version and bootfiles for developer's bootstrapping
2005    convenience.
2006    
2007       110.43.2 -- NEW BOOTFILES
2008    
2009    ----------------------------------------------------------------------
2010    Name: Matthias Blume (blume (at) tti - c (dot) org)
2011    Date: 2003/09/19 15:55:00 CDT
2012    Tag: blume-20030919-cmdir
2013    Description:
2014    
2015    1. new-install.sh -> install.sh
2016    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
2017    3. tweaked installer so that another name instead of .cm can be chosen
2018       at install time (by setting the CM_DIR_ARC environment variable
2019       during installation); once installation is complete, the name is
2020       fixed
2021    
2022    ----------------------------------------------------------------------
2023    Name: Matthias Blume (blume (at) tti - c (dot) org)
2024    Date: 2003/09/18 16:00:00 CDT
2025    Tag: blume-20030918-110_43_1
2026    Description:
2027    
2028    Made a new interim version and bootfiles for developer's bootstrapping
2029    convenience.
2030    
2031       110.43.1 -- NEW BOOTFILES
2032    
2033    ----------------------------------------------------------------------
2034    Name: Matthias Blume (blume (at) tti - c (dot) org)
2035    Date: 2003/09/18 15:20:00 CDT
2036    Tag: blume-20030918-misc
2037    Description:
2038    
2039    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
2040       this seems to be controversial at the moment) and used those in
2041       Posix.ProcEnv.times.
2042    
2043    2. Added Time.{from,to}Nanoseconds to Time.
2044    
2045    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
2046       For example, fromLargeInt never needs to look at more than 3 "big
2047       digits" to get its 53 bits of precision.
2048    
2049    ----------------------------------------------------------------------
2050    Name: Matthias Blume (blume (at) tti - c (dot) org)
2051    Date: 2003/09/17 16:30:00 CDT
2052    Tag: blume-20030917-real32-slices
2053    Description:
2054    
2055    Added an entry to the primitive environment
2056    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
2057    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
2058    
2059    Removed some of the "magic" constants in real64.sml and replaced them
2060    with code that generates these values from their corresponding
2061    integer counterparts.
2062    
2063    Made all(?) the slice-related changes to the Basis and made everything
2064    compile again...
2065    
2066    ----------------------------------------------------------------------
2067    Name: Matthias Blume (blume (at) tti - c (dot) org)
2068    Date: 2003/09/15 17:45:00 CDT
2069    Tag: blume-20030915-rbase
2070    Description:
2071    
2072    Fixed bug in Real.fromLargeInt.
2073    
2074    ----------------------------------------------------------------------
2075    Name: Matthias Blume (blume (at) tti - c (dot) org)
2076    Date: 2003/09/13 18:11:00 CDT
2077    Tag: blume-20030913-libinstall
2078    Description:
2079    
2080    Minor bugfix in config/libinstall (set anchor with path to
2081    standalone tool after installing it, otherwise libraries that
2082    need ml-lex or ml-yacc won't compile the first time the installer
2083    runs).
2084    
2085    ----------------------------------------------------------------------
2086    Name: Matthias Blume (blume (at) tti - c (dot) org)
2087    Date: 2003/09/12 11:45:00 CDT
2088    Tag: blume-20030912-various
2089    Description:
2090    
2091     - fixed bug in Real.toLargeInt
2092     - fixed bug in Posix.ProcEnv.times
2093     - changed inputLine functions to return an option
2094     - minor installer improvements / bugfixes
2095     - changed default @SMLalloc parameter for x86/celeron to 64k
2096    
2097    ----------------------------------------------------------------------
2098    Name: Matthias Blume (blume (at) tti - c (dot) org)
2099    Date: 2003/09/09 22:00:00 CDT
2100    Tag: Release_110_43
2101    Description:
2102    
2103        New working release 110.43.  New bootfiles.
2104    
2105    ----------------------------------------------------------------------
2106    Name: Matthias Blume (blume (at) tti - c (dot) org)
2107    Date: 2003/09/09 19:20:00 CDT
2108    Tag: blume-20030909-installer
2109    Description:
2110    
2111    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
2112    Modified config/install.bat to take advantage of it.  Also modified
2113    config/install.sh (and called it config/new-install.sh) to take advantage
2114    of it on Unix systems.  (The SML code is (supposed to be) platform-
2115    independent.)
2116    
2117    The installer can now install everything under Win32
2118    as well as under *nix as long as it compiles.
2119    
2120    Other changes:
2121    
2122      - made CML compile again under Win32
2123      - made eXene compile under Win32 (by providing a fake structure UnixSock
2124        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
2125      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
2126        same as type int;  under Win32 it isn't
2127      - fixed some slice-related problems in the win32-specific parts of CML
2128      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
2129        (under Win32, the volume name is case-insensitive, and the
2130        OS.Path code compares volume names for equality)
2131    
2132    ----------------------------------------------------------------------
2133    Name: Matthias Blume (blume (at) tti - c (dot) org)
2134    Date: 2003/09/08 11:55:00 CDT
2135    Tag: blume-20030908-fullpath
2136    Description:
2137    
2138    Made Win32 version of OS.FileSys.fullPath return current directory
2139    when given an empty string.  This is what the spec says, and incidentally,
2140    CM depends on it.  (CM otherwise goes into an infinite loop in certain
2141    cases when presented with the name of a non-existing .cm file.)
2142    
2143    ----------------------------------------------------------------------
2144    Name: Matthias Blume (blume (at) tti - c (dot) org)
2145    Date: 2003/09/04 16:30:00 CDT
2146    Tag: blume-20030905-slices-etc
2147    Description:
2148    
2149      1. Changed interface to vectors and arrays in Basis to match
2150         (draft) Basis spec.
2151      2. Added signatures and implementations of slices according to
2152         Basis spec.
2153      3. Edited source code throughout the system to make it compile again
2154         under 1. and 2.  (In some cases code had to be added to have it
2155         match the new signatures.)
2156      4. MLRISC should be backward-compatible: the copies of the originals
2157         of files that needed to change under 3. were retained, the .cm files
2158         check the compiler version number and use old versions when
2159         appropriate.
2160      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
2161         dirstream -> string option (in accordance with Basis spec).
2162      6. When generating code that counts lines, ml-lex used function
2163         CharVector.foldli, taking advantage of its old interface.
2164         This has been replaced with the corresponding code from
2165         CharVectorSlice.  (html-lex must be re-lexed!)
2166      7. BitArray in smlnj-lib/Util has been extended/modified to match the
2167         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
2168      8. Removed temporary additions (fromInternal, toInternal) from the
2169         (now obsolete) IntInf in smlnj-lib/Util.
2170      9. Cleaned up structure Byte.
2171     10. Added localOffset, scan, and fromString to Date (according to spec).
2172         Cleaned/corrected implementation of Date.
2173         (Still need to check for correctness; implement better canonicalizeDate.)
2174     11. Added "scan" to signature IEEE_REAL.
2175     12. Some improvements to IntInf [in particular: efficiency-hack for
2176         mod and rem when second operand is 2 (for parity checks).]
2177     13. Changed representation of type Time.time, using a single IntInf.int
2178         value counting microseconds.  This considerably simplified the
2179         implementation of structure Time.  We now support negative time
2180         values; scan and fromString handle signs.
2181     14. Functor PrimIO now takes two additional arguments (VectorSlice and
2182         ArraySlice).
2183    
2184    ----------------------------------------------------------------------
2185    Name: Matthias Blume (blume (at) tti - c (dot) org)
2186    Date: 2003/08/28 17:00:00 CDT
2187    Tag: blume-20030828-intinf
2188    Description:
2189    
2190    This is a major update which comes with a version number bump
2191    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
2192    and an implementation of IntInf in the Basis.
2193    
2194    There are a fairly large number of related changes and updates throughout
2195    the system:
2196    
2197    Basis:
2198            - Implemented IntInf.
2199            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
2200            - Added some missing Real64 operations, most notably Real.toLargeInt.
2201            - Added FixedInt as a synonym for Int32.
2202    
2203    compiler:
2204            * Added support for a built-in intinf type.
2205              - literals
2206              - pattern matching
2207              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
2208              - overloading on literals and operations
2209    
2210            This required adding a primitive type intinf, some additional
2211            primops, and implementations for several non-trivial intinf
2212            operations in Core.  (The intinf type is completely abstract
2213            to the compiler; all operations get delegated back to the Core.)
2214    
2215            * Intinf equality is handled by polyequal.  However, the compiler
2216            does not print its usual warning in this case (since polyequal
2217            is the right thing to do there).
2218    
2219            * Improved the organization of structure InlineT.
2220    
2221            * A word about conversion primops:
2222            If conversions involving intinf do not cancel out during
2223            CPS contract, then the compiler must insert calls to Core functions.
2224            Since all core access must be resolved already during the FLINT
2225            translate phase, it would be too late a the time of CPS contract
2226            to add new Core calls.  For this reason, conversion primops
2227            for intinf carry two arguments:  1. the numeric argument that
2228            they are supposed to convert, and 2. the Core function that
2229            can help with this conversion if necessary.  If CPS contract
2230            eliminates a primop, then the associated Core function becomes
2231            dead and goes away.  Intinf conversion primops that do not get
2232            eliminated by CPS contract get rewritten into calls of their
2233            core functions by a separate, new phase.
2234    
2235    interactive system:
2236            - Control.Print.intinfDepth controls max length of intinf constants
2237            being printed. (Analogous to Control.Print.stringDepth.)
2238            - Cleanup in printutil and pputil: got rid of unused stuff and
2239            duplicates; replaced some of the code with code that makes better
2240            use of library functionality.
2241    
2242    CM:
2243            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
2244            are no longer silent.
2245    
2246    CKIT:
2247            Fixed mismatched uses of Int32 and LargeInt.  I always decided
2248            in favor of LargeInt -- which is now the same as IntInf.
2249            CKIT-knowledgable people should check whether this is what's
2250            intended and otherwise change things back to using Int32 or
2251            FixedInt.
2252    
2253    Throughout the code:
2254            Started using IntInf.int literals and built-in operations
2255            (e.g., comparison with 0) where this seems appropriate.
2256    
2257    
2258    ----------------------------------------------------------------------
2259    Name: Dave MacQueen (dbm@cs.uchicago.edu)
2260    Date: 2003/08/13 11:36:00 CDT
2261    Tag: dbm-20030813-mcz-merge1
2262    Description:
2263    
2264    Merging changes from the mcz-branch development branch into trunk.
2265    These changes involve replacement of the emulated old prettyprinter
2266    interface with direct use of the SML/NJ Lib PP library, and fixing
2267    of a couple of bugs (895, 1186) relating to error messages.  A new
2268    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
2269    has been added.
2270    
2271    ----------------------------------------------------------------------
2272    Name: Matthias Blume (blume (at) tti - c (dot) org)
2273    Date: 2003/08/11 15:45:00 CDT
2274    Tag: blume-20030811-windows
2275    Description:
2276    
2277    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
2278    
2279        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
2280    
2281    This patch restores SML/NJ's ability to run under win32.  There are a
2282    number of changes, including fixes for several bugs that had gone
2283    unnoticed until now:
2284    
2285      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
2286        trying to run the win32 version from within cygwin.)
2287    
2288      - There are a number of simple .bat scripts that substitute for their
2289        corresponding Unix shell-scripts.  (See below.)
2290    
2291      - The internals of ml-build have been modified slightly.  The main
2292        difference is that instead of calling ".link-sml" (or link-sml.bat)
2293        using OS.Process.system, the ML process delegates this task back
2294        to the script.  Otherwise problems arise in mixed environments such
2295        as Cygwin where scripts look and work like Unix scripts, but
2296        where OS.Process.system cannot run them.
2297    
2298      - In CM, the srcpath pickler used native pathname syntax -- which
2299        is incorrect in the case of cross-compilation.  The new pickle format
2300        is independent of platform-specific naming conventions.
2301    
2302      - Path configuration files (such as lib/pathconfig) can now choose
2303        between native and standard syntax.  Placing a line of the form
2304    
2305            standard!
2306    
2307        into the file causes all subsequent paths to be interpreted using
2308        CM standard pathname syntax (= Unix conventions); a line
2309    
2310            native!
2311    
2312        switches back to native style.  This was needed so that
2313        path config files can be written portably, see src/system/pathconfig.
2314    
2315      - Runtime system:
2316    
2317          - win32-filesys.c:  get_file_time and set_file_time now
2318            access modification time, not creation time.
2319    
2320          - I/O code made aware of new array representation.
2321    
2322          - Bug fixes in X86.prim.masm.
2323    
2324       - src/system/makeml made aware of win32. (For use under cygwin
2325            and other Unix-environments for windows.)
2326    
2327       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
2328         which caused BinIO.inputAll to fail consistently.
2329    
2330    .bat scripts:
2331    
2332       Windows .bat scripts assume that SMLNJ_HOME is defined.
2333    
2334       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
2335         applications (sml, ml-yacc, ml-lex).
2336       - ml-build.bat: analogous to ml-build.
2337       - config\install.bat:  Analogous to config/install.sh.  This requires
2338         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
2339         (nmake etc. must be on the path, and vcvars32 must have been run.)
2340         Moreover, sources for ml-lex and ml-yacc need to exist under src,
2341         and the bootfile hierarchy must have been unpacked under
2342         sml.boot.x86-win32.
2343         The script is very primitive and does a poor job at error checking.
2344         It only installs the base system, ml-lex, and ml-yacc.  No other
2345         libraries are being installed (i.e., you get only those that
2346         are part of the compiler.)
2347       - link-sml.bat:  analogous to .link-sml, but not currently used
2348    
2349    Unrelated bug fixes:
2350    
2351      - ml-nlffigen now exports structures ST_* corresponding to incomplete
2352        types.
2353      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
2354        otherwise.)
2355    
2356    ----------------------------------------------------------------------
2357    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
2358    Date: 2003/06/17
2359    Tag: macqueen-20030617-bug895
2360    Description:
2361    
2362    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
2363    Tag will be used for new development branch (mcz-branch) for
2364    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
2365    
2366    ----------------------------------------------------------------------
2367    Name: Matthias Blume (blume (at) tti - c (dot) org)
2368    Date: 2003/05/27 16:55:00 CDT
2369    Tag: blume-20030527-polyeq
2370    Description:
2371    
2372    Tried to eliminated most cases of polymorphic equality.
2373    
2374    ----------------------------------------------------------------------
2375    Name: Matthias Blume (blume (at) tti - c (dot) org)
2376    Date: 2003/05/21 17:45:00 CDT
2377    Tag: blume-20030517-complete
2378    Description:
2379    
2380    Two changes:
2381    
2382      1. Added a flag for controlling whether non-exhaustive bindings will
2383         be treated as errors (default is false).
2384      2. Cleaned up the *entire* source tree so that CMB.make goes through
2385         without a single non-exhaustive match- or bind warning.
2386    
2387    ----------------------------------------------------------------------
2388    Name: Matthias Blume (blume (at) tti - c (dot) org)
2389    Date: 2003/05/17 10:20:00 CDT
2390    Tag: blume-20030517-absyn
2391    Description:
2392    
2393    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
2394    
2395       This mainly affects the quality of error messages.  However, some
2396       of the code is now more straightforward than before. (Treatment of
2397       the above four constructs in translate.sml is much simpler than
2398       the "macro-expansion" that was going on before.  Plus, the mach-
2399       compiler no longer gets invoked just to be able to compile an
2400       if-expression.)
2401    
2402    2. The ErrorMsg.Error exception is now caught and absorbed by the
2403       interactive loop.
2404    
2405    ----------------------------------------------------------------------
2406    Name: Allen Leung
2407    Date: 2003/05/16 13:05:00 CDT
2408    Tag: leunga-20030516-cygwin-runtime
2409    Description:
2410    
2411        Ported the runtime system to cygwin, which uses the unix
2412    x86-unix bin files.  Missing/buggy features:
2413    
2414       o getnetbyname, getnetbyaddr: these functions seem to be missing in
2415         the Cygwin library.
2416       o Ctrl-C handling may be flaky.
2417       o Windows system calls and Windows I/O are not supported.
2418    
2419       A new set of binfiles is located at:
2420    
2421         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
2422    
2423       This is only needed for bootstrapping the cygwin version of smlnj.
2424       Other x86 versions can use the existing binfiles.
2425    
2426    ----------------------------------------------------------------------
2427    Name: Matthias Blume
2428    Date: 2003/04/08 15:42:00 CDT
2429    Tag: blume-20030408-listpair
2430    Description:
2431    
2432    1. Added a target 'mlrisc' to installer.
2433    
2434    2. Added missing elements to structure ListPair.
2435    
2436    ----------------------------------------------------------------------
2437    Name: Allen Leung
2438    Date: 2003/01/07 10:40:00 EST
2439    Tag: leunga-20030107-int-rem
2440    Description:
2441    
2442    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
2443    The arguments to the SUBL instruction were swapped.
2444    
2445    ----------------------------------------------------------------------
2446    Name: Matthias Blume
2447    Date: 2002/12/12 16:25:00 EST
2448    Tag: blume-20021212-risc-ra
2449    Description:
2450    
2451    Fixed a serious bug in the rewrite code for FP spilling/reloading that
2452    sent the RA into an infinite loop when floating point registers get
2453    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
2454    110.37 and 110.38.)
2455    There was another set of potential problems related to the handling of
2456    MLRISC annotations (but those did not yet cause real problems, apparently).
2457    
2458    ----------------------------------------------------------------------
2459    Name: Matthias Blume
2460    Date: 2002/12/06 22:40:00 EST
2461    Tag: blume-20021206-cm-fileid
2462    Description:
2463    
2464    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
2465    This fixes the problem of CM getting confused by files that suddenly
2466    change their identity (e.g., by getting unlinked and recreated by some
2467    text editor such as vi).  There might be a better/cheaper/cleaner way
2468    of doing this, but for now this will have to do.
2469    
2470    ----------------------------------------------------------------------
2471    Name: Matthias Blume
2472    Date: 2002/10/28 09:50:00 EST
2473    Tag: blume-20021028-typecheck
2474    Description:
2475    
2476    Exported structure Typecheck from $smlnj/viscomp/core.cm.
2477    
2478    ----------------------------------------------------------------------
2479    Name: Matthias Blume
2480    Date: 2002/10/17 09:10:00 EDT
2481    Tag: Release_110_42
2482    Description:
2483    
2484    In good old tradition, there has been a slight hiccup so that we have
2485    to patch 110.42 after the fact.   The old release tag has been replaced
2486    (see below).
2487    
2488    The change solves a problem with two competing approaches the
2489    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
2490    each other's way.
2491    
2492    This change only affects the runtime system code and the installer script.
2493    (No new bootfiles.)
2494    
2495    ----------------------------------------------------------------------
2496    Name: Matthias Blume
2497    Date: 2002/10/16 12:00:00 EDT
2498    Tag: Release_110_42_removed
2499    Description:
2500    
2501    New working release.  New bootfiles.
2502    
2503    ----------------------------------------------------------------------
2504    Name: Matthias Blume
2505    Date: 2002/10/10 13:10:00 EDT
2506    Tag: blume-20021010-ppc-divs
2507    Description:
2508    
2509    The mltree operator DIVS must be implemented with an overflow check on
2510    the PPC because the hardware indicates divide-by-zero using "overflow" as
2511    well.
2512    
2513    ----------------------------------------------------------------------
2514    Name: Matthias Blume
2515    Date: 2002/07/23 11:45:00
2516    Tag: blume-20020723-smlnj-home
2517    Description:
2518    
2519    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
2520    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
2521    CM_PATHCONFIG is also set), the path configuration file is assumed
2522    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
2523    the entire tree to some other place and everything will "just work".
2524    
2525    (Companion commands such as ml-build and ml-makedepend also sense this
2526    variable.)
2527    
2528    ----------------------------------------------------------------------
2529    Name: Matthias Blume
2530    Date: 2002/07/12 21:19:00 EDT
2531    Tag: blume-20020712-liveness
2532    Description:
2533    
2534    Exported two useful "step" functions from liveness module (MLRISC).
2535    
2536    ----------------------------------------------------------------------
2537    Name: Matthias Blume
2538    Date: 2002/07/05 16:00 EDT
2539    Tag: Release_110_41
2540    Description:
2541    
2542    New working release.  New bootfiles.
2543    
2544    ----------------------------------------------------------------------
2545    Name: Matthias Blume
2546    Date: 2002/07/05 10:25:00 EDT
2547    Tag: blume-20020705-btimp
2548    Description:
2549    
2550    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
2551    other clients can set up backtracing support.
2552    
2553    ----------------------------------------------------------------------
2554    Name: Matthias Blume
2555    Date: 2002/06/25 17:23:00 EDT
2556    Tag: blume-20020625-fpmax
2557    Description:
2558    
2559    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
2560    case.  (The sense of the isNaN test was reversed -- which made min and
2561    max always return their first argument.)
2562    
2563    ----------------------------------------------------------------------
2564    Name: Matthias Blume
2565    Date: 2002/06/11
2566    Tag: blume-20020611-unixpath
2567    Description:
2568    
2569    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
2570    
2571    ----------------------------------------------------------------------
2572    Name: Matthias Blume
2573    Date: 2002/06/10 16:35:00 EDT
2574    Tag: blume-20020610-ieeereal
2575    Description:
2576    
2577    I back-ported my implementation of IEEEReal.fromString from the
2578    idlbasis-devel branch so that we can test it.
2579    
2580    Another small change is that ppDec tries to give more information
2581    than just "<sig>" in the case of functors.  However, this code is
2582    broken in some mysterious way if the functor's body's signature
2583    has not been declared by ascription but gets inferred from the
2584    implementation.  This needs fixing...
2585    
2586    ----------------------------------------------------------------------
2587    Name: Matthias Blume
2588    Date: 2002/05/31
2589    Tag: blume-20020531-btrace-mode
2590    Description:
2591    
2592    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
2593    the wayside when I switched over to using Controls everywhere.)
2594    
2595    ----------------------------------------------------------------------
2596    Name: Lal George
2597    Date: 2002/05/23 12:21:40 EDT
2598    Tag: george-20020523-visual-labels
2599    Description:
2600    
2601    Labels are now displayed in the graphical output to make
2602    the fall-through and target blocks obvious.
2603    
2604    ----------------------------------------------------------------------
2605    Name: Matthias Blume
2606    Date: 2002/05/22 11:03:00 EDT
2607    Tag: blume-20020522-shrink
2608    Description:
2609    
2610    John tweaked yesterday's fix for 1131 to handle an out-of-memory
2611    situation that comes up when allocating huge arrays.
2612    
2613    ----------------------------------------------------------------------
2614    Name: Matthias Blume
2615    Date: 2002/05/21 16:00:00 EDT
2616    Tag: Release_110_40
2617    Description:
2618    
2619    New working release (110.40).  New bootfiles.
2620    
2621    [Also: John Reppy fixed GC bug 1131.]
2622    
2623    ----------------------------------------------------------------------
2624    Name: Matthias Blume
2625    Date: 2002/05/21 12:35:00 EDT
2626    Tag: blume-20020521-cmdoc
2627    Description:
2628    
2629    CM documentation update.
2630    
2631    ----------------------------------------------------------------------
2632    Name: Matthias Blume
2633    Date: 2002/05/21 10:55:00 EDT
2634    Tag: blume-20020521-misc
2635    Description:
2636    
2637    - John tweaked runtime to be silent on heap export (except when
2638      GC messages are on).
2639    - I added a few more things (cross-compiling versions of CMB) to
2640      config/preloads (as suggestions).
2641    
2642    ----------------------------------------------------------------------
2643    Name: Matthias Blume
2644    Date: 2002/05/20 22:25:00 EDT
2645    Tag: blume-20020520-controls
2646    Description:
2647    
2648    - Added ControlUtil structure to control-lib.cm.
2649    - Use it throughout.
2650    - Used Controls facility to define MLRISC controls (as opposed to
2651      registering MLRISC control ref cells with Controls after the fact)
2652    - Fixed messed-up controls priorities.
2653    
2654    * Removed again all the stuff from config/preloads that one wouldn't
2655      be able to preload at the time the initial heap image is built.
2656      (Many libraries, e.g., CML, do not exist yet at this time.  The
2657       only libraries that can be preloaded via config/preloads are those
2658       that come bundled with the bootfiles.)
2659    
2660    ----------------------------------------------------------------------
2661    Name: Matthias Blume
2662    Date: 2002/05/20 10:59:00 EDT
2663    Tag: blume-20020520-preloads
2664    Description:
2665    
2666    Added a lot of commented-out suggestions for things to be included
2667    in config/preloads.
2668    
2669    ----------------------------------------------------------------------
2670    Name: Allen Leung
2671    Date: 2002/05/18 14:20:00 EDT
2672    Tag: leunga-20020518-mdl
2673    Description:
2674    
2675    o Made the mdl tool stuff compile and run again.
2676    o I've disabled all the stuff that depends on RTL specifications; they
2677      are all badly broken anyway.
2678    
2679    ----------------------------------------------------------------------
2680    Name: Matthias Blume
2681    Date: 2002/05/17 16:49:00 EDT
2682    Tag: blume-20020517-controls
2683    Description:
2684    
2685    0. John Reppy made several modifications to the SML/NJ library.
2686       In particular, there is a shiny new controls-lib.cm.
2687    
2688    1. Pushed new controls interface through compiler so that everything
2689       compiles again.
2690    
2691    2. Added FormatComb and FORMAT_COMB to the CML version of the
2692       SML/NJ library (so that CML compiles again).
2693    
2694    3. Modified init scripts because XXX_DEFAULT environment variables
2695       are no longer with us.  (Boot-time initialization is now done
2696       using the same environment variables that are also used for
2697       startup-time initialization of controls.)
2698    
2699    ----------------------------------------------------------------------
2700    Name: Lal George
2701    Date: 2002/05/15 09:20:10 EDT
2702    Tag: george-20020515-pseudo-op-decls
2703    Description:
2704    
2705            All pseudo-ops emitted before the first segment declaration
2706    such as TEXT, DATA, and BSS directives are assumed to be global
2707    declarations and are emitted first in the assembly file. This is
2708    useful in a number of situations where one has pseudo-ops that are not
2709    specific to any segment, and also works around the constraint that one
2710    cannot have client pseudo-ops in the TEXT segment.
2711    
2712            Because no segment is associated with these declarations it is
2713    an error to allocate any space or objects before the first segment
2714    directive and an exception will be raised. However, we cannot make
2715    this check for client pseudo-ops.
2716    
2717            These top level declarations are a field in the CFG graph_info.
2718    In theory you can continue to add to this field after the CFG has been
2719    built -- provided you know what you are doing;-)
2720    
2721    ----------------------------------------------------------------------
2722    Name: Matthias Blume
2723    Date: 2002/05/13 16:40:00 EDT
2724    Tag: blume-20020513-pp-etc
2725    Description:
2726    
2727    A few minor bugfixes:
2728    
2729      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
2730        (Bogus printouts for redefined bindings still occur.  Compiler
2731        bug should no longer occur now.  We need to redo the prettyprinter
2732        from scratch.)
2733    
2734      - CM pathname printer now also adds escape sequences for ( and )
2735    
2736      - commend and docu fixes for ml-nlffi
2737    
2738    ----------------------------------------------------------------------
2739    Name: Matthias Blume
2740    Date: 2002/05/10 16:40:00 EDT
2741    Tag: blume-20020510-erg-textio
2742    Description:
2743    
2744    Applied the following bugfix provided by Emden Gansner:
2745    
2746        Output is corrupted when outputSubstr is used rather than output.
2747    
2748        The problem occurs when a substring
2749    
2750            ss = (s, dataStart, dataLen)
2751    
2752        where dataStart > 0, fills a stream buffer with avail bytes left.
2753        avail bytes of s, starting at index dataStart, are copied into the
2754        buffer, the buffer is flushed, and then the remaining dataLen-avail
2755        bytes of ss are copied into the beginning of the buffer. Instead of
2756        starting this copy at index dataStart+avail in s, the current code
2757        starts the copy at index avail.
2758    
2759        Fix:
2760        In text-io-fn.sml, change line 695 from
2761             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2762        to
2763             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2764    
2765    ----------------------------------------------------------------------
2766    Name: Matthias Blume
2767    Date: 2002/04/12 13:55:00 EDT
2768    Tag: blume-20020412-assyntax
2769    Description:
2770    
2771    1. Grabbed newer assyntax.h from the XFree86 project.
2772    2. Fiddled with how to compile X86.prim.asm without warnings.
2773    3. (Very) Minor cleanup in CM.
2774    
2775    ----------------------------------------------------------------------
2776    Name: Matthias Blume
2777    Date: 2002/04/01 (no joke!) 17:07:00 EST
2778    Tag: blume-20020401-x86div
2779    Description:
2780    
2781    Added full support for div/mod/rem/quot on the x86, using the machine
2782    instruction's two results (without clumsily recomputing the remainder)
2783    directly where appropriate.
2784    
2785    Some more extensive power-of-two support was added to the x86 instruction
2786    selector (avoiding expensive divs, mods, and muls where they can be
2787    replaced with cheaper shifts and masks).  However, this sort of thing
2788    ought to be done earlier, e.g., within the CPS optimizer so that
2789    all architectures benefit from it.
2790    
2791    The compiler compiles to a fixed point, but changes might be somewhat
2792    fragile nevertheless.  Please, report any strange things that you might
2793    see wrt. div/mod/quot/rem...
2794    
2795    ----------------------------------------------------------------------
2796    Name: Matthias Blume
2797    Date: 2002/03/29 17:22:00
2798    Tag: blume-20020329-div
2799    Description:
2800    
2801    Fixed my broken div/mod logic.  Unfortunately, this means that the
2802    inline code for div/mod now has one more comparison than before.
2803    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2804    The problem was with quotient = 0, because that alone does not tell
2805    us which way the rounding went.  One then has to look at whether
2806    remainder and divisor have the same sign...  :(
2807    
2808    Anyway, I replaced the bootfiles with fresh ones...
2809    
2810    ----------------------------------------------------------------------
2811    Name: Matthias Blume
2812    Date: 2002/03/29 14:10:00 EST
2813    Tag: blume-20020329-inlprims
2814    Description:
2815    
2816    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2817    
2818    Primops have changed. This means that the bin/boot-file formats have
2819    changed as well.
2820    
2821    To make sure that there is no confusion, I made a new version.
2822    
2823    
2824    CHANGES:
2825    
2826    * removed REMT from mltree (remainder should never overflow).
2827    
2828    * added primops to deal with divisions of all flavors to the frontend
2829    
2830    * handled these primops all the way through so they map to their respective
2831      MLRISC support
2832    
2833    * used these primops in the implementation of Int, Int32, Word, Word32
2834    
2835    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2836    
2837    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2838    
2839    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2840      floating point
2841    
2842    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2843      and Real (but Real.abs maps to a separate floating-point-only primop)
2844    
2845    
2846    TODO items:
2847    
2848    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2849      instructions because the machine instruction encoder cannot handle
2850      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2851      REM in the first place, and REM is supported by the x86 machine coder.)
2852    
2853    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2854      selection phase.  (The two can be streamlined because the hardware
2855      delivers both quotient and remainder at the same time anyway.)
2856    
2857    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2858      (Currently the behavior is inconsistent both across architectures and
2859      wrt. the draft Basis spec.)
2860    
2861    * Word8 should eventually be handled natively, too.
2862    
2863    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2864      as if there currently is no execution path that could trigger it in
2865      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2866      hold for things like multiplication and division.)
2867    
2868    ----------------------------------------------------------------------
2869    Name: Matthias Blume
2870    Date: 2002/03/27 16:27:00 EST
2871    Tag: blume-20020327-mlrisc-divisions
2872    Description:
2873    
2874    Added support for all four division operations (ML's div, mod, quot,
2875    and rem) to MLRISC.  In the course of doing so, I also rationalized
2876    the naming (no more annoying switch-around of DIV and QUOT), by
2877    parameterizing the operation by div_rounding_mode (which can be either
2878    DIV_TO_ZERO or DIV_TO_NEGINF).
2879    
2880    The generic MLTreeGen functor takes care of compiling all four
2881    operations down to only round-to-zero div.
2882    
2883    Missing pieces:
2884    
2885      * Doing something smarter than relying on MLTreeGen on architectures
2886        like, e.g., the x86 where hardware division delivers both quotient and
2887        remainder at the same time.  With this, the implementation of the
2888        round-to-neginf operations could be further streamlined.
2889    
2890      * Remove inlining support for div/mod/rem from the frontend and replace it
2891        with primops that get carried through to the backend.  Do this for all
2892        int and word types.
2893    
2894    ----------------------------------------------------------------------
2895    Name: Matthias Blume
2896    Date: 2002/03/25 17:25:00 EST
2897    Tag: blume-20020325-divmod
2898    Description:
2899    
2900    I improved (hopefully without breaking them) the implementation of Int.div,
2901    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2902    advantage of the following observations:
2903    
2904      Let  q = x quot y      r = x rem y
2905           d = x div  y      m = x mod y
2906    
2907    where "quot" is the round-to-zero version of integer division that
2908    hardware usually provides.  Then we have:
2909    
2910         r = x - q * y        where neither the * nor the - will overflow
2911         d = if q >= 0 orelse x = q * y then q else q - 1
2912                              where neither the * nor the - will overflow
2913         m = if q >= 0 orelse r = 0 then r else r + y
2914                              where the + will not overflow
2915    
2916    This results in substantial simplification of the generated code.
2917    The following table shows the number of CFG nodes and edges generated
2918    for
2919            fun f (x, y) = x OPER y
2920            (* with OPER \in div, mod, quot, rem *)
2921    
2922    
2923        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2924        --------------------------------------------------------
2925         div |         24 |         39 |         12 |         16
2926         mod |         41 |         71 |         12 |         16
2927        quot |          8 |         10 |          8 |         10
2928         rem |         10 |         14 |          8 |         10
2929    
2930    
2931    ----------------------------------------------------------------------
2932    Name: Matthias Blume
2933    Date: 2002/03/25 22:06:00 EST
2934    Tag: blume-20020325-cprotobug
2935    Description:
2936    
2937    Fixed a bug in cproto (c prototype decoder).
2938    
2939    ----------------------------------------------------------------------
2940    Name: Matthias Blume
2941    Date: 2002/03/25 16:00:00 EST
2942    Tag: blume-20020325-raw-primops
2943    Description:
2944    
2945    I did some cleanup to Allen's new primop code and
2946    replaced yesterday's bootfiles with new ones.
2947    (But they are stored in the same place.)
2948    
2949    ----------------------------------------------------------------------
2950    Name: Matthias Blume
2951    Date: 2002/03/24 22:40:00 EST
2952    Tag: blume-20020324-bootfiles
2953    Description:
2954    
2955    Made the bootfiles that Allen asked for.
2956    
2957    ----------------------------------------------------------------------
2958    Name: Allen Leung
2959    Date: 2002/03/23 15:50:00 EST
2960    Tag: leunga-20020323-flint-cps-rcc-primops
2961    Description:
2962    
2963      1. Changes to FLINT primops:
2964    
2965        (* make a call to a C-function;
2966         * The primop carries C function prototype information and specifies
2967         * which of its (ML-) arguments are floating point. C prototype
2968         * information is for use by the backend, ML information is for
2969         * use by the CPS converter. *)
2970      | RAW_CCALL of { c_proto: CTypes.c_proto,
2971                       ml_args: ccall_type list,
2972                       ml_res_opt: ccall_type option,
2973                       reentrant : bool
2974                     } option
2975       (* Allocate uninitialized storage on the heap.
2976        * The record is meant to hold short-lived C objects, i.e., they
2977        * are not ML pointers.  With the tag, the representation is
2978        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2979        *)
2980      | RAW_RECORD of {tag:bool,sz:int}
2981      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2982    
2983      2.  These CPS primops are now overloaded:
2984    
2985           rawload of {kind:numkind}
2986           rawstore of {kind:numkind}
2987    
2988          The one argument form is:
2989    
2990             rawload {kind} address
2991    
2992          The two argument form is:
2993    
2994             rawload {kind} [ml object, byte-offset]
2995    
2996      3. RAW_CCALL/RCC now takes two extra arguments:
2997    
2998         a. The first is whether the C call is reentrant, i.e., whether
2999            ML state should be saved and restored.
3000         b. The second argument is a string argument specifying the name of
3001            library and the C function.
3002    
3003         These things are currently not handled in the code generator, yet.
3004    
3005      4. In CProto,
3006    
3007         An encoding type of "bool" means "ml object" and is mapped into
3008         C prototype of PTR.  Note that "bool" is different than "string",
3009         even though "string" is also mapped into PTR, because "bool"
3010         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
3011    
3012      5. Pickler/unpicker
3013    
3014         Changed to handle RAW_RECORD and newest RAW_CCALL
3015    
3016      6. MLRiscGen,
3017    
3018         1. Changed to handle the new rawload/rawstore/rawrecord operators.
3019         2. Code for handling C Calls has been moved to a new module CPSCCalls,
3020            in the file CodeGen/cpscompile/cps-c-calls.sml
3021    
3022      7. Added the conditional move operator
3023    
3024             condmove of branch
3025    
3026         to cps.  Generation of this is still buggy so it is currently
3027         disabled.
3028    
3029    ----------------------------------------------------------------------
3030    Name: Lal George
3031    Date: 2002/03/22 14:18:25 EST
3032    Tag: george-20020322-cps-branch-prob
3033    Description:
3034    
3035    Implemented the Ball-Larus branch prediction-heuristics, and
3036    incorporated graphical viewers for control flow graphs.
3037    
3038    Ball-Larus Heuristics:
3039    ---------------------
3040    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
3041    
3042    By design it uses the Dempster-Shafer theory for combining
3043    probabilities.  For example, in the function:
3044    
3045        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
3046    
3047    the ball-larus heuristics predicts that the n=0 is unlikely
3048    (OH-heuristic), and the 'then' branch is unlikely because of the
3049    RH-heuristic -- giving the 'then' branch an even lower combined
3050    probability using the Dempster-Shafer theory.
3051    
3052    Finally, John Reppy's loop analysis in MLRISC, further lowers the
3053    probability of the 'then' branch because of the loop in the else
3054    branch.
3055    
3056    
3057    Graphical Viewing:
3058    ------------------
3059    I merely plugged in Allen's graphical viewers into the compiler. The
3060    additional code is not much. At the top level, saying:
3061    
3062            Control.MLRISC.getFlag "cfg-graphical-view" := true;
3063    
3064    will display the graphical view of the control flow graph just before
3065    back-patching.  daVinci must be in your path for this to work. If
3066    daVinci is not available, then the default viewer can be changed
3067    using:
3068    
3069            Control.MLRISC.getString "viewer"
3070    
3071    which can be set to "dot" or "vcg" for the corresponding viewers. Of
3072    course, these viewers must be in your path.
3073    
3074    The above will display the compilation unit at the level of clusters,
3075    many of which are small, boring, and un-interesting. Also setting:
3076    
3077            Control.MLRISC.getInt "cfg-graphical-view_size"
3078    
3079    will display clusters that are larger than the value set by the above.
3080    
3081    
3082    ----------------------------------------------------------------------
3083    Name: Matthias Blume
3084    Date: 2002/03/21 22:20:00 EST
3085    Tag: blume-20020321-kmp-bugfix
3086    Description:
3087    
3088    Changed the interface to the KMP routine in PreString and fixed
3089    a minor bug in one place where it was used.
3090    
3091    ----------------------------------------------------------------------
3092    Name: Allen Leung
3093    Date: 2002/03/21 20:30:00 EST
3094    Tag: leunga-20020321-cfg
3095    Description:
3096    
3097      Fixed a potential problem in cfg edge splitting.
3098    
3099    ----------------------------------------------------------------------
3100    Name: Allen Leung
3101    Date: 2002/03/21 17:15:00 EST
3102    Tag: leunga-20020321-x86-fp-cfg
3103    Description:
3104    
3105      1. Recoded the buggy parts of x86-fp.
3106    
3107         a. All the block reordering code has been removed.
3108            We now depend on the block placement phases to do this work.
3109    
3110         b. Critical edge splitting code has been simplified and moved into the
3111            CFG modules, as where they belong.
3112    
3113         Both of these were quite buggy and complex.  The code is now much, much
3114         simpler.
3115    
3116      2. X86 backend.
3117    
3118         a. Added instructions for 64-bit support.  Instruction selection for
3119            64-bit has not been committed, however, since that
3120            requires changes to MLTREE which haven't been approved by
3121            Lal and John.
3122    
3123         b. Added support for FUCOMI and FUCOMIP when generating code for
3124            PentiumPro and above.  We only generate these instructions in
3125            the fast-fp mode.
3126    
3127         c. Added cases for JP and JNP in X86FreqProps.
3128    
3129      3. CFG
3130    
3131         CFG now has a bunch of methods for edge splitting and merging.
3132    
3133      4. Machine description.
3134    
3135         John's simplification of MLTREE_BASIS.fcond broke a few machine
3136         description things:
3137    
3138         rtl-build.{sig,sml} and hppa.mdl fixed.
3139    
3140         NOTE: the machine description stuff in the repository is still broken.
3141               Again, I can't put my fixes in because that involves
3142               changes to MLTREE.
3143    
3144    ----------------------------------------------------------------------
3145    Name: Matthias Blume
3146    Date: 2002/03/20 15:55:00 EST
3147    Tag: blume-20020320-kmp
3148    Description:
3149    
3150    Implemented Knuth-Morris-Pratt string matching in PreString and used
3151    it for String.isSubstring, Substring.isSubstring, and
3152    Substring.position.
3153    
3154    (Might need some stress-testing.  Simple examples worked fine.)
3155    
3156    ----------------------------------------------------------------------
3157    Name: Matthias Blume
3158    Date: 2002/03/19 16:37:00 EST
3159    Tag: blume-20020319-witnesses
3160    Description:
3161    
3162    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
3163    
3164    This implements a generic mechanism for changing constness qualifiers
3165    anywhere within big C types without resorting to outright "casts".
3166    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
3167    modify the constness at the outermost level.)
3168    The implementation of "convert" is based on the idea of "witness"
3169    values -- values that are not used by the operation but whose types
3170    "testify" to their applicability.  On the implementation side, "convert"
3171    is simply a projection (returning its second curried argument).  With
3172    cross-module inlining, it should not result in any machine code being
3173    generated.
3174    
3175    ----------------------------------------------------------------------
3176    Name: Matthias Blume
3177    Date: 2002/03/15 16:40:00 EST
3178    Tag: blume-20020315-basis
3179    Description:
3180    
3181    Provided (preliminary?) implementations for
3182    
3183      {String,Substring}.{concatWith,isSuffix,isSubstring}
3184    
3185    and
3186    
3187      Substring.full
3188    
3189    Those are in the Basis spec but they were missing in SML/NJ.
3190    
3191    ----------------------------------------------------------------------
3192    Name: Matthias Blume
3193    Date: 2002/03/14 21:30:00 EST
3194    Tag: blume-20020314-controls
3195    Description:
3196    
3197    Controls:
3198    ---------
3199    
3200    1. Factored out the recently-added Controls : CONTROLS stuff and put
3201       it into its own library $/controls-lib.cm.  The source tree for
3202       this is under src/smlnj-lib/Controls.
3203    
3204    2. Changed the names of types and functions in this interface, so they
3205       make a bit more "sense":
3206    
3207          module -> registry
3208          'a registry -> 'a group
3209    
3210    3. The interface now deals in ref cells only.  The getter/setter interface
3211       is (mostly) gone.
3212    
3213    4. Added a function that lets one register an already-existing ref cell.
3214    
3215    5. Made the corresponding modifications to the rest of the code so that
3216       everything compiles again.
3217    
3218    6. Changed the implementation of Controls.MLRISC back to something closer
3219       to the original.  In particular, this module (and therefore MLRISC)
3220       does not depend on Controls.  There now is some link-time code in
3221       int-sys.sml that registers the MLRISC controls with the Controls
3222       module.
3223    
3224    CM:
3225    ---
3226    
3227      * One can now specify the lambda-split aggressiveness in init.cmi.
3228    
3229    ----------------------------------------------------------------------
3230    Name: Allen Leung
3231    Date: 2002/03/13 17:30:00 EST
3232    Tag: leunga-20020313-x86-fp-unary
3233    Description:
3234    
3235    Bug fix for:
3236    
3237    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
3238    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
3239    > - fun f(x,(y,z)) = Real.~ y;
3240    > [autoloading]
3241    > [autoloading done]
3242    >       fchsl   (%eax), 184(%esp)
3243    > Error: MLRisc bug: X86MCEmitter.emitInstr
3244    >
3245    > uncaught exception Error
3246    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
3247    
3248    The problem was that the code generator did not generate any fp registers
3249    in this case, and the ra didn't know that it needed to run the X86FP phase to
3250    translate the pseudo fp instruction.   This only happened with unary fp
3251    operators in certain situations.
3252    
3253    ----------------------------------------------------------------------
3254    Name: Matthias Blume
3255    Date: 2002/03/13 14:00:00 EST
3256    Tag: blume-20020313-overload-etc
3257    Description:
3258    
3259    1. Added _overload as a synonym for overload for backward compatibility.
3260       (Control.overloadKW must be true for either version to be accepted.)
3261    
3262    2. Fixed bug in install script that caused more things to be installed
3263       than what was requested in config/targets.
3264    
3265    3. Made CM aware of the (_)overload construct so that autoloading
3266       works.
3267    
3268    ----------------------------------------------------------------------
3269    Name: Matthias Blume
3270    Date: 2002/03/12 22:03:00 EST
3271    Tag: blume-20020312-url
3272    Description:
3273    
3274    Forgot to update BOOT and srcarchiveurl.
3275    
3276    ----------------------------------------------------------------------
3277    Name: Matthias Blume
3278    Date: 2002/03/12 17:30:00 EST
3279    Tag: blume-20020312-version110392
3280    Description:
3281    
3282    Yet another version number bump (because of small changes to the
3283    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
3284    
3285    Changes:
3286    
3287      The new pid generation scheme described a few weeks ago was overly
3288      complicated.  I implemented a new mechanism that is simpler and
3289      provides a bit more "stability":  Once CM has seen a compilation
3290      unit, it keeps its identity constant (as long as you do not delete
3291      those crucial CM/GUID/* files).  This means that when you change
3292      an interface, compile, then go back to the old interface, and
3293      compile again, you arrive at the original pid.
3294    
3295      There now also is a mechanism that instructs CM to use the plain
3296      environment hash as a module's pid (effectively making its GUID
3297      the empty string).  For this, "noguid" must be specified as an
3298      option to the .sml file in question within its .cm file.
3299      This is most useful for code that is being generated by tools such
3300      as ml-nlffigen (because during development programmers tend to
3301      erase the tool's entire output directory tree including CM's cached
3302      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
3303      revert to the old, broken behavior of SML/NJ, but in specific cases
3304      where there is no danger of interface confusion, its use is ok
3305      (I think).
3306    
3307      ml-nlffigen by default generates "noguid" annotations.  They can be
3308      turned off by specifying -guid in its command line.
3309    
3310    ----------------------------------------------------------------------
3311    Name: Lal George
3312    Date: 2002/03/12 12 14:42:36 EST
3313    Tag: george-20020312-frequency-computation
3314    Description:
3315    
3316    Integrated jump chaining and static block frequency into the
3317    compiler. More details and numbers later.
3318    
3319    ----------------------------------------------------------------------
3320    Name: Lal George
3321    Date: 2002/03/11 11 22:38:53 EST
3322    Tag: george-20020311-jump-chain-elim
3323    Description:
3324    
3325    Tested the jump chain elimination on all architectures (except the
3326    hppa).  This is on by default right now and is profitable for the
3327    alpha and x86, however, it may not be profitable for the sparc and ppc
3328    when compiling the compiler.
3329    
3330    The gc test will typically jump to a label at the end of the cluster,
3331    where there is another jump to an external cluster containing the actual
3332    code to invoke gc. This is to allow factoring of common gc invocation
3333    sequences. That is to say, we generate:
3334    
3335            f:
3336               testgc
3337               ja   L1      % jump if above to L1
3338    
3339            L1:
3340               jmp L2
3341    
3342    
3343    After jump chain elimination the 'ja L1' instructions is converted to
3344    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
3345    up being implemented in their long form (if L2 is far away) using:
3346    
3347            jbe     L3      % jump if below or equal to L3
3348            jmp     L2
3349         L3:
3350            ...
3351    
3352    
3353    For large compilation units L2  may be far away.
3354    
3355    
3356    ----------------------------------------------------------------------
3357    Name: Matthias Blume
3358    Date: 2002/03/11 13:30:00 EST
3359    Tag: blume-20020311-mltreeeval
3360    Description:
3361    
3362    A functor parameter was missing.
3363    
3364    ----------------------------------------------------------------------
3365    Name: Allen Leung
3366    Date: 2002/03/11 10:30:00 EST
3367    Tag: leunga-20020311-runtime-string0
3368    Description:
3369    
3370       The representation of the empty string now points to a
3371    legal null terminated C string instead of unit.  It is now possible
3372    to convert an ML string into C string with InlineT.CharVector.getData.
3373    This compiles into one single machine instruction.
3374    
3375    ----------------------------------------------------------------------
3376    Name: Allen Leung
3377    Date: 2002/03/10 23:55:00 EST
3378    Tag: leunga-20020310-x86-call
3379    Description:
3380    
3381       Added machine generation for CALL instruction (relative displacement mode)
3382    
3383    ----------------------------------------------------------------------
3384    Name: Matthias Blume
3385    Date: 2002/03/08 16:05:00
3386    Tag: blume-20020308-entrypoints
3387    Description:
3388    
3389    Version number bumped to 110.39.1.  NEW BOOTFILES!
3390    
3391    Entrypoints: non-zero offset into a code object where execution should begin.
3392    
3393    - Added the notion of an entrypoint to CodeObj.
3394    - Added reading/writing of entrypoint info to Binfile.
3395    - Made runtime system bootloader aware of entrypoints.
3396    - Use the address of the label of the first function given to mlriscGen
3397      as the entrypoint.  This address is currently always 0, but it will
3398      not be 0 once we turn on block placement.
3399    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
3400      with entry points) from mlriscGen.
3401    
3402    ----------------------------------------------------------------------
3403    Name: Allen Leung
3404    Date: 2002/03/07 20:45:00 EST
3405    Tag: leunga-20020307-x86-cmov
3406    Description:
3407    
3408       Bug fixes for CMOVcc on x86.
3409    
3410       1. Added machine code generation for CMOVcc
3411       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
3412       3. CMOVcc cannot have an immediate operand as argument.
3413    
3414    ----------------------------------------------------------------------
3415    Name: Matthias Blume
3416    Date: 2002/03/07 16:15:00 EST
3417    Tag: blume-20020307-controls
3418    Description:
3419    
3420    This is a very large but mostly boring patch which makes (almost)
3421    every tuneable compiler knob (i.e., pretty much everything under
3422    Control.* plus a few other things) configurable via both the command
3423    line and environment variables in the style CM did its configuration
3424    until now.
3425    
3426    Try starting sml with '-h' (or, if you are brave, '-H')
3427    
3428    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
3429    implements the underlying generic mechanism.
3430    
3431    The interface to some of the existing such facilities has changed somewhat.
3432    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
3433    (The getFoo interface is still there for backward-compatibility, but its
3434    use is deprecated.)
3435    
3436    The ml-build script passes -Cxxx=yyy command-line arguments through so
3437    that one can now twiddle the compiler settings when using this "batch"
3438    compiler.
3439    
3440    TODO items:
3441    
3442    We should go through and throw out all controls that are no longer
3443    connected to anything.  Moreover, we should go through and provide
3444    meaningful (and correct!) documentation strings for those controls
3445    that still are connected.
3446    
3447    Currently, multiple calls to Controls.new are accepted (only the first
3448    has any effect).  Eventually we should make sure that every control
3449    is being made (via Controls.new) exactly once.  Future access can then
3450    be done using Controls.acc.
3451    
3452    Finally, it would probably be a good idea to use the getter-setter
3453    interface to controls rather than ref cells.  For the time being, both
3454    styles are provided by the Controls module, but getter-setter pairs are
3455    better if thread-safety is of any concern because they can be wrapped.
3456    
3457    *****************************************
3458    
3459    One bug fix: The function blockPlacement in three of the MLRISC
3460    backpatch files used to be hard-wired to one of two possibilities at
3461    link time (according to the value of the placementFlag).  But (I
3462    think) it should rather sense the flag every time.
3463    
3464    *****************************************
3465    
3466    Other assorted changes (by other people who did not supply a HISTORY entry):
3467    
3468    1. the cross-module inliner now works much better (Monnier)
3469    2. representation of weights, frequencies, and probabilities in MLRISC
3470       changed in preparation of using those for weighted block placement
3471       (Reppy, George)
3472    
3473    ----------------------------------------------------------------------
3474    Name: Lal George
3475    Date: 2002/03/07 14:44:24 EST 2002
3476    Tag: george-20020307-weighted-block-placement
3477    
3478    Tested the weighted block placement optimization on all architectures
3479    (except the hppa) using AMPL to generate the block and edge frequencies.
3480    Changes were required in the machine properties to correctly
3481    categorize trap instructions. There is an MLRISC flag
3482    "weighted-block-placement" that can be used to enable weighted block
3483    placement, but this will be ineffective without block/edge
3484    frequencies (coming soon).
3485    
3486    
3487    ----------------------------------------------------------------------
3488    Name: Lal George
3489    Date: 2002/03/05 17:24:48 EST
3490    Tag: george-20020305-linkage-cluster
3491    
3492    In order to support the block placement optimization, a new cluster
3493    is generated as the very first cluster (called the linkage cluster).
3494    It contains a single jump to the 'real' entry point for the compilation
3495    unit. Block placement has no effect on the linkage cluster itself, but
3496    all the other clusters  have full freedom in the manner in which they
3497    reorder blocks or functions.
3498    
3499    On the x86 the typical linkage code that is generated is:
3500       ----------------------
3501            .align 2
3502       L0:
3503            addl    $L1-L0, 72(%esp)
3504            jmp     L1
3505    
3506    
3507            .align  2
3508       L1:
3509       ----------------------
3510    
3511    72(%esp) is the memory location for the stdlink register. This
3512    must contain the address of the CPS function being called. In the
3513    above example, it contains the address of  L0; before
3514    calling L1 (the real entry point for the compilation unit), it
3515    must contain the address for L1, and hence
3516    
3517            addl $L1-L0, 72(%esp)
3518    
3519    I have tested this on all architectures except the hppa.The increase
3520    in code size is of course negligible
3521    
3522    ----------------------------------------------------------------------
3523    Name: Allen Leung
3524    Date: 2002/03/03 13:20:00 EST
3525    Tag: leunga-20020303-mlrisc-tools
3526    
3527      Added #[ ... ] expressions to mlrisc tools
3528    
3529    ----------------------------------------------------------------------
3530    Name: Matthias Blume
3531    Date: 2002/02/27 12:29:00 EST
3532    Tag: blume-20020227-cdebug
3533    Description:
3534    
3535    - made types in structure C and C_Debug to be equal
3536    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
3537    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
3538    
3539    ----------------------------------------------------------------------
3540    Name: Matthias Blume
3541    Date: 2002/02/26 12:00:00 EST
3542    Tag: blume-20020226-ffi
3543    Description:
3544    
3545    1. Fixed a minor bug in CM's "noweb" tool:
3546       If numbering is turned off, then truly don't number (i.e., do not
3547       supply the -L option to noweb).  The previous behavior was to supply
3548       -L'' -- which caused noweb to use the "default" line numbering scheme.
3549       Thanks to Chris Richards for pointing this out (and supplying the fix).
3550    
3551    2. Once again, I reworked some aspects of the FFI:
3552    
3553       A. The incomplete/complete type business:
3554    
3555       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
3556         gone!
3557       - ML types representing an incomplete type are now *equal* to
3558         ML types representing their corresponding complete types (just like
3559         in C).  This is still safe because ml-nlffigen will not generate
3560         RTTI for incomplete types, nor will it generate functions that
3561         require access to such RTTI.   But when ML code generated from both
3562         incomplete and complete versions of the C type meet, the ML types
3563         are trivially interoperable.
3564    
3565         NOTE:  These changes restore the full generality of the translation
3566         (which was previously lost when I eliminated functorization)!
3567    
3568       B. Enum types:
3569    
3570       - Structure C now has a type constructor "enum" that is similar to
3571         how the "su" constructor works.  However, "enum" is not a phantom
3572         type because each "T enum" has values (and is isomorphic to
3573         MLRep.Signed.int).
3574       - There are generic access operations for enum objects (using
3575         MLRep.Signed.int).
3576       - ml-nlffigen will generate a structure E_foo for each "enum foo".
3577         * The structure contains the definition of type "mlrep" (the ML-side
3578         representation type of the enum).  Normally, mlrep is the same
3579         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
3580         then mlrep will be defined as a datatype -- thus facilitating
3581         pattern matching on mlrep values.
3582         ("-ec" will be suppressed if there are duplicate values in an
3583          enumeration.)
3584         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
3585         will be generated for each C enum constant xxx.
3586         * Conversion functions m2i and i2m convert between mlrep and
3587         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
3588         * Coversion functions c and ml convert between mlrep and "tag enum".
3589         * Access functions (get/set) fetch and store mlrep values.
3590       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
3591         enumerations are merged into one single enumeration represented by
3592         structure E_'.
3593    
3594    ----------------------------------------------------------------------
3595    Name: Allen Leung
3596    Date: 2002/02/25 04:45:00 EST
3597    Tag: leunga-20020225-cps-spill
3598    
3599    This is a new implementation of the CPS spill phase.
3600    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
3601    In case of problems, replace it with the old file spill.sml
3602    
3603    The current compiler runs into some serious performance problems when
3604    constructing a large record.  This can happen when we try to compile a
3605    structure with many items.  Even a very simple structure like the following
3606    makes the compiler slow down.
3607    
3608        structure Foo = struct
3609           val x_1 = 0w1 : Word32.int
3610           val x_2 = 0w2 : Word32.int
3611           val x_3 = 0w3 : Word32.int
3612           ...
3613           val x_N = 0wN : Word32.int
3614        end
3615    
3616    The following table shows the compile time, from N=1000 to N=4000,
3617    with the old compiler:
3618    
3619    N
3620    1000   CPS 100 spill                           0.04u  0.00s  0.00g
3621           MLRISC ra                               0.06u  0.00s  0.05g
3622              (spills = 0 reloads = 0)
3623           TOTAL                                   0.63u  0.07s  0.21g
3624    
3625    1100   CPS 100 spill                           8.25u  0.32s  0.64g
3626           MLRISC ra                               5.68u  0.59s  3.93g
3627              (spills = 0 reloads = 0)
3628           TOTAL                                   14.71u  0.99s  4.81g
3629    
3630    1500   CPS 100 spill                           58.55u  2.34s  1.74g
3631           MLRISC ra                               5.54u  0.65s  3.91g
3632              (spills = 543 reloads = 1082)
3633           TOTAL                                   65.40u  3.13s  6.00g
3634    
3635    2000   CPS 100 spill                           126.69u  4.84s  3.08g
3636           MLRISC ra                               0.80u  0.10s  0.55g
3637              (spills = 42 reloads = 84)
3638           TOTAL                                   129.42u  5.10s  4.13g
3639    
3640    3000   CPS 100 spill                           675.59u  19.03s  11.64g
3641           MLRISC ra                               2.69u  0.27s  1.38g
3642              (spills = 62 reloads = 124)
3643           TOTAL                                   682.48u  19.61s  13.99g
3644    
3645    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
3646           MLRISC ra                               4.96u  0.27s  2.72g
3647              (spills = 85 reloads = 170)
3648           TOTAL                                   2375.26u  57.21s  48.00g
3649    
3650    As you can see the old cps spill module suffers from some serious
3651    performance problem.  But since I cannot decipher the old code fully,
3652    instead of patching the problems up, I'm reimplementing it
3653    with a different algorithm.  The new code is more modular,
3654    smaller when compiled, and substantially faster
3655    (O(n log n) time and O(n) space).  Timing of the new spill module:
3656    
3657    4000  CPS 100 spill                           0.02u  0.00s  0.00g
3658          MLRISC ra                               0.25u  0.02s  0.15g
3659             (spills=1 reloads=3)
3660          TOTAL                                   7.74u  0.34s  1.62g
3661    
3662    Implementation details:
3663    
3664    As far as I can tell, the purpose of the CPS spill module is to make sure the
3665    number of live variables at any program point (the bandwidth)
3666    does not exceed a certain limit, which is determined by the
3667    size of the spill area.
3668    
3669    When the bandwidth is too large, we decrease the register pressure by
3670    packing live variables into spill records.  How we achieve this is
3671    completely different than what we did in the old code.
3672    
3673    First, there is something about the MLRiscGen code generator
3674    that we should be aware of:
3675    
3676    o MLRiscGen performs code motion!
3677    
3678       In particular, it will move floating point computations and
3679       address computations involving only the heap pointer to
3680       their use sites (if there is only a single use).
3681       What this means is that if we have a CPS record construction
3682       statement
3683    
3684           RECORD(k,vl,w,e)
3685    
3686       we should never count the new record address w as live if w
3687       has only one use (which is often the case).
3688    
3689       We should do something similar to floating point, but the transformation
3690       there is much more complex, so I won't deal with that.
3691    
3692    Secondly, there are now two new cps primops at our disposal:
3693    
3694     1. rawrecord of record_kind option
3695        This pure operator allocates some uninitialized storage from the heap.
3696        There are two forms:
3697    
3698         rawrecord NONE [INT n]  allocates a tagless record of length n
3699         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
3700                                     and initializes the tag.
3701    
3702     2. rawupdate of cty
3703          rawupdate cty (v,i,x)
3704          Assigns to x to the ith component of record v.
3705          The storelist is not updated.
3706    
3707    We use these new primops for both spilling and increment record construction.
3708    
3709     1. Spilling.
3710    
3711        This is implemented with a linear scan algorithm (but generalized
3712        to trees).  The algorithm will create a single spill record at the
3713        beginning of the cps function and use rawupdate to spill to it,
3714        and SELECT or SELp to reload from it.  So both spills and reloads
3715        are fine-grain operations.  In contrast, in the old algorithm
3716        "spills" have to be bundled together in records.
3717    
3718        Ideally, we should sink the spill record construction to where
3719        it is needed.  We can even split the spill record into multiple ones
3720        at the places where they are needed.  But CPS is not a good
3721        representation for global code motion, so I'll keep it simple and
3722        am not attempting this.
3723    
3724     2. Incremental record construction (aka record splitting).
3725    
3726        Long records with many component values which are simulatenously live
3727        (recall that single use record addresses are not considered to
3728         be live) are constructed with rawrecord and rawupdate.
3729        We allocate space on the heap with rawrecord first, then gradually
3730        fill it in with rawupdate.  This is the technique suggested to me
3731        by Matthias.
3732    
3733        Some restrictions on when this is applicable:
3734        1. It is not a VECTOR record.  The code generator currently does not handle
3735           this case. VECTOR record uses double indirection like arrays.
3736        2. All the record component values are defined in the same "basic block"
3737           as the record constructor.  This is to prevent speculative
3738           record construction.
3739    
3740    ----------------------------------------------------------------------
3741    Name: Allen Leung
3742    Date: 2002/02/22 01:02:00 EST
3743    Tag: leunga-20020222-mlrisc-tools
3744    
3745    Minor bug fixes in the parser and rewriter
3746    
3747    ----------------------------------------------------------------------
3748    Name: Allen Leung
3749    Date: 2002/02/21 20:20:00 EST
3750    Tag: leunga-20020221-peephole
3751    
3752    Regenerated the peephole files.  Some contained typos in the specification
3753    and some didn't compile because of pretty printing bugs in the old version
3754    of 'nowhere'.
3755    
3756    ----------------------------------------------------------------------
3757    Name: Allen Leung
3758    Date: 2002/02/19 20:20:00 EST
3759    Tag: leunga-20020219-mlrisc-tools
3760    Description:
3761    
3762       Minor bug fixes to the mlrisc-tools library:
3763    
3764       1.  Fixed up parsing colon suffixed keywords
3765       2.  Added the ability to shut the error messages up
3766       3.  Reimplemented the pretty printer and fixed up/improved
3767           the pretty printing of handle and -> types.
3768       4.  Fixed up generation of literal symbols in the nowhere tool.
3769       5.  Added some SML keywords to to sml.sty
3770    
3771    ----------------------------------------------------------------------
3772    Name: Matthias Blume
3773    Date: 2002/02/19 16:20:00 EST
3774    Tag: blume-20020219-cmffi
3775    Description:
3776    
3777    A wild mix of changes, some minor, some major:
3778    
3779    * All C FFI-related libraries are now anchored under $c:
3780        $/c.cm      --> $c/c.cm
3781        $/c-int.cm  --> $c/internals/c-int.cm
3782        $/memory.cm --> $c/memory/memory.cm
3783    
3784    * "make" tool (in CM) now treats its argument pathname slightly
3785      differently:
3786        1. If the native expansion is an absolute name, then before invoking
3787           the "make" command on it, CM will apply OS.Path.mkRelative
3788           (with relativeTo = OS.FileSys.getDir()) to it.
3789        2. The argument will be passed through to subsequent phases of CM
3790           processing without "going native".  In particular, if the argument
3791           was an anchored path, then "make" will not lose track of that anchor.
3792    
3793    * Compiler backends now "know" their respective C calling conventions
3794      instead of having to be told about it by ml-nlffigen.  This relieves
3795      ml-nlffigen from one of its burdens.
3796    
3797    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3798    
3799    * Export C_DEBUG and C_Debug from $c/c.cm.
3800    
3801    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3802      subtyping relationship between incomplete pointers and their complete
3803      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3804      with the convention of instantiating 'o with ('t, 'c) obj whenever
3805      the pointer target type is complete.  In the incomplete case, 'o
3806      will be instantiated with some "'c iobj" -- a type obtained by
3807      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3808    
3809      Operations that work on both incomplete and complete pointer types are
3810      typed as taking an 'o ptr while operations that require the target to
3811      be known are typed as taking some ('t, 'c) obj ptr.
3812    
3813      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3814      where void is an eqtype without any values.  This makes it possible
3815      to work on voidptr values using functions meant to operate on light
3816      incomplete pointers.
3817    
3818    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3819      been vastly simplified.
3820    
3821    ----------------------------------------------------------------------
3822    Name: Matthias Blume
3823    Date: 2002/02/19 10:48:00 EST
3824    Tag: blume-20020219-pqfix
3825    Description:
3826    
3827    Applied Chris Okasaki's bug fix for priority queues.
3828    
3829    ----------------------------------------------------------------------
3830    Name: Matthias Blume
3831    Date: 2002/02/15 17:05:00
3832    Tag: Release_110_39
3833    Description:
3834    
3835    Last-minute retagging is becoming a tradition... :-(
3836    
3837    This is the working release 110.39.
3838    
3839    ----------------------------------------------------------------------
3840    Name: Matthias Blume
3841    Date: 2002/02/15 16:00:00 EST
3842    Tag: Release_110_39-orig
3843    Description:
3844    
3845    Working release 110.39.  New bootfiles.
3846    
3847    (Update: There was a small bug in the installer so it wouldn't work
3848    with all shells.  So I retagged. -Matthias)
3849    
3850    ----------------------------------------------------------------------
3851    Name: Matthias Blume
3852    Date: 2002/02/15 14:17:00 EST
3853    Tag: blume-20020215-showbindings
3854    Description:
3855    
3856    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3857    the latter can be useful for exploring what bindings are available at
3858    the interactive prompt.  (The first function returns only the list
3859    of symbols that are really bound, the second prints those but also the
3860    ones that CM's autoloading mechanism knows about.)
3861    
3862    ----------------------------------------------------------------------
3863    Name: Matthias Blume
3864    Date: 2002/02/15 12:08:00 EST
3865    Tag: blume-20020215-iptrs
3866    Description:
3867    
3868    Two improvements to ml-nlffigen:
3869    
3870      1. Write files only if they do not exist or if their current contents
3871         do not coincide with what's being written.  (That is, avoid messing
3872         with the time stamps unless absolutely necessary.)
3873    
3874      2. Implement a "repository" mechanism for generated files related
3875         to "incomplete pointer types".   See the README file for details.
3876    
3877    ----------------------------------------------------------------------
3878    Name: Matthias Blume
3879    Date: 2002/02/14 11:50:00 EST
3880    Tag: blume-20020214-quote
3881    Description:
3882    
3883    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3884    because of the new and improved tag generation scheme.  (Thanks to Allen
3885    Leung for pointing it out.)
3886    
3887    ----------------------------------------------------------------------
3888    Name: Lal George
3889    Date: 2002/02/14 09:55:27 EST 2002
3890    Tag: george-20020214-isabelle-bug
3891    Description:
3892    
3893    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3894    of Isabelle on the x86.
3895    
3896    From Allen:
3897    -----------
3898     I've found the problem:
3899    
3900         in ra-core.sml, I use the counter "blocked" to keep track of the
3901         true number of elements in the freeze queue.  When the counter goes
3902         to zero, I skip examining the queue.  But I've messed up the
3903         bookkeeping in combine():
3904    
3905             else ();
3906             case !ucol of
3907               PSEUDO => (if !cntv > 0 then
3908                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3909                                        ^^^^^^^^^^^^^^^^^^^^^^^
3910                      moveu := mergeMoveList(!movev, !moveu)
3911                     )
3912                  else ();
3913    
3914         combine() is called to coalesce two nodes u and v.
3915         I think I was thinking that if the move counts of u and v are both
3916         greater than zero then after they are coalesced then one node is
3917         removed from the freeze queue.  Apparently I was thinking that
3918         both u and v are of low degree, but that's clearly not necessarily true.
3919    
3920    
3921    02/12/2002:
3922        Here's the patch.  HOL now compiles.
3923    
3924        I don't know how this impact on performance (compile
3925        time or runtime).  This bug caused the RA (especially on the x86)
3926        to go thru the potential spill phase when there are still nodes on the
3927        freeze queue.
3928    
3929    
3930    
3931    
3932    ----------------------------------------------------------------------
3933    Name: Matthias Blume
3934    Date: 2002/02/13 22:40:00 EST
3935    Tag: blume-20020213-fptr-rtti
3936    Description:
3937    
3938    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3939    updates.
3940    
3941    ----------------------------------------------------------------------
3942    Name: Matthias Blume
3943    Date: 2002/02/13 16:41:00 EST
3944    Tag: blume-20020213-cmlpq
3945    Description:
3946    
3947    Added new priority queue export symbols (which have just been added to
3948    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3949    not compile and the installer would choke.)
3950    
3951    ----------------------------------------------------------------------
3952    Name: Matthias Blume
3953    Date: 2002/02/13 16:15:00 EST
3954    Tag: blume-20020213-various
3955    Description:
3956    
3957    1. More tweaks to ml-nlffigen:
3958    
3959       - better internal datastructures (resulting in slight speedup)
3960       - "-match" option requires exact match
3961       - "localized" gensym counters (untagged structs/unions nested within
3962         other structs/unions or within typedefs get a fresh counter; their
3963         tag will be prefixed by a concatenation of their parents' tags)
3964       - bug fixes (related to calculation of transitive closure of types
3965         to be included in the output)
3966    
3967    2. Minor Basis updates:
3968    
3969       - added implementations for List.collate and Option.app
3970    
3971    ----------------------------------------------------------------------
3972    Name: Matthias Blume
3973    Date: 2002/02/11 15:55:00 EST
3974    Tag: blume-20020211-gensym
3975    Description:
3976    
3977    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3978    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3979    names (ML structure names that correspond to unnamed C structs, unions,
3980    and enums), so that separate runs of ml-nlffigen do not clash.
3981    
3982    ----------------------------------------------------------------------
3983    Name: Matthias Blume
3984    Date: 2002/02/11 12:05:00 EST
3985    Tag: blume-20020211-gensml
3986    Description:
3987    
3988    A quick fix for a problem with GenSML (in the pgraph-util library):
3989    Make generation of toplevel "local" optional.  (Strictly speaking,
3990    signature definitions within "local" are not legal SML.)
3991    
3992    Other than that: updates to INSTALL and cm/TODO.
3993    
3994    ----------------------------------------------------------------------
3995    Name: Matthias Blume
3996    Date: 2002/02/08 15:00:00 EST
3997    Tag: blume-20020208-uniquepid
3998    Description:
3999    
4000    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
4001    
4002    1. The installer (config/install.sh) has gotten smarter:
4003    
4004         - Configuration options are a bit easier to specify now
4005           (in config/targets).
4006         - Bug in recognizing .tar.bz2 files fixed.
4007         - Installer automatically resolves dependencies between
4008           configuration options (e.g., if you ask for eXene, you will
4009           also get cml -- regardless whether you asked for it or not).
4010         - Installer can run in "quieter mode" by setting the environment
4011           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
4012           "completely silent", though.
4013         - Build HashCons library as part of smlnj-lib.
4014    
4015    2. A new scheme for assigning persistent identifiers to compilation
4016       units (and, by extension, to types etc.) has been put into place.
4017       This fixes a long-standing bug where types and even dynamic values
4018       can get internally confused, thereby compromising type safety
4019       (abstraction) and dynamic correctness.  See
4020    
4021         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
4022    
4023       for an example of how things could go wrong until now.
4024    
4025       The downside of the new scheme is that pids are not quite as
4026       persistent as they used to be: CM will generate a fresh pid
4027       for every compilation unit that it thinks it sees for the first
4028       time.  That means that if you compile starting from a clean, fresh
4029       source tree at two different times, you end up with different
4030       binaries.
4031    
4032       Cutoff recompilation, however, has not been compromised because
4033       CM keeps pid information in special caches between runs.
4034    
4035    ----------------------------------------------------------------------
4036    Name: Lal George
4037    Date: 2002/02/07 15:34:13 EST 2002
4038    Tag: <none>
4039    Description:
4040    
4041    Compilers that generate assembly code may produce  global labels
4042    whose value is resolved at link time. The various peephole optimization
4043    modules did not take this in account.
4044    
4045    TODO. The Labels.addrOf function should really return an option
4046    type so that clients are forced to deal with this issue, rather
4047    than an exception being raised.
4048    
4049    ----------------------------------------------------------------------
4050    Name: Lal George
4051    Date: 2002/02/06 13:55:02 EST
4052    Tag: george-20020206-ra-breakup
4053    Description:
4054    
4055    1. A bug fix from Allen.
4056    
4057        A typo causes extra fstp %st(0)'s to be generated at compensation
4058        edges, which might cause stack underflow traps at runtime.  This
4059        occurs in fft where there are extraneous fstps right before the 'into'
4060        trap instruction (in this case they are harmless since none of the
4061        integers overflow.)
4062    
4063    2. Pulled out various utility modules that were embedded in the modules
4064       of the register allocator. I need these modules for other purposes, but
4065       they are not complete enough to put into a library (just yet).
4066    ----------------------------------------------------------------------
4067    Name: Matthias Blume
4068    Date: 2002/01/31 16:05:00 EST
4069    Tag: blume-20020131-sparc-ccalls
4070    Description:
4071    
4072    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
4073       of extra stack space by mistake.  Fixed.
4074    
4075    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
4076    
4077    ----------------------------------------------------------------------
4078    Name: Allen Leung
4079    Date: 2002/01/30
4080    Tag: leunga-20020130-nowhere-bug-fix
4081    Description:
4082    
4083       MLRISC bug fixes:
4084       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
4085       2. MachineInt.fromString was negating its value.
4086    
4087    ----------------------------------------------------------------------
4088    Name: Matthias Blume
4089    Date: 2002/01/29
4090    Tag: blume-20020129-INSTALL
4091    Description:
4092    
4093    - Added somewhat detailed installation instructions (file INSTALL).
4094    - Fixed curl-detection bug in config/install.sh.
4095    - It is now possible to select the URL getter using the URLGETTER
4096      environment variable:
4097    
4098          not set / "unknown"      --> automatic detection (script tries wget,