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 570, Wed Mar 8 17:30:13 2000 UTC revision 1873, Sun Nov 20 05:32:27 2005 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:  Name: Matthias Blume (http://www.tti-c.org/blume/)
17  Date:  Date: 2005/11/19 23:35:00 CST
18    Tag: blume-20051119-fish
19    Description:
20    
21    Fixed a problem in config/install.sh where it tries to "fish" the
22    name of the CM metadata directory from the wrong place (because
23    the physical location of basis.cm has changed).  Also, corrected the
24    path anchor for $/html-lib.cm.  (Thanks to M. Fluet for pointing out
25    these problems.)
26    
27    ----------------------------------------------------------------------
28    Name: Matthias Blume (http://www.tti-c.org/blume/)
29    Date: 2005/11/16 14:30:00 CST
30    Tag: Release_110_57
31    Description:
32    
33    New working version (110.57).  NEW BOOTFILES!
34    
35    ----------------------------------------------------------------------
36    Name: Matthias Blume (http://www.tti-c.org/blume/)
37    Date: 2005/11/16 10:25:00 CST
38    Tag: blume-22051116-btrace
39    Description:
40    
41    Fixed problem with bogus exception message when using back-trace
42    facility.
43    
44    ----------------------------------------------------------------------
45    Name: Matthias Blume (http://www.tti-c.org/blume/)
46    Date: 2005/11/15 22:35:00 CST
47    Tag: blume-22051115-array2copy
48    Description:
49    
50    Added simple implementation of Array2.copy.  (Warning: mostly untested.)
51    
52    ----------------------------------------------------------------------
53    Name: Matthias Blume (http://www.tti-c.org/blume/)
54    Date: 2005/11/15 21:35:00 CST
55    Tag: blume-22051115-versiontool
56    Description:
57    
58    Reversed change to src/system/smlnj/internals/versiontool.cm.  This
59    file gets loaded as a tool -- by the equivalent of CM.make during the
60    run of CMB.make.  Thus, CMB's path configuration is meaningless for
61    it.  Instead, it has the status of "user code", so it should use
62    $/basis.cm to refer to the Basis library.  (At least that's true for
63    the purpose of bootstrapping the previous change.  In the future it
64    might make sense to have versiontool.cm refer to
65    $smlnj/basis/basis.cm, i.e., the version of the Basis that the
66    compiler itself uses.)
67    
68    Also patched src/system/testml to have it activate those extra anchor
69    bindings in config/extrapathconfig.
70    
71    ----------------------------------------------------------------------
72    Name: Matthias Blume (http://www.tti-c.org/blume/)
73    Date: 2005/11/15 16:35:00 CST
74    Tag: blume-22051115-libraries
75    Description:
76    
77    This change affects the way the following libraries are tied into
78    the system:
79    
80       $/basis.cm
81       $/smlnj-lib.cm
82       $/pp-lib.cm
83       $/controls-lib.cm
84       $/html-lib.cm
85       $/ml-yacc-lib.cm
86    
87    These libraries are now internally (as seen from the source code
88    of the implementation itself) known by the following names:
89    
90       $smlnj/basis/basis.cm
91       $smlnj/smlnj-lib/smlnj-lib.cm
92       $smlnj/smlnj-lib/pp-lib.cm
93       $smlnj/smlnj-lib/controls-lib.cm
94       $smlnj/smlnj-lib/html-lib.cm
95       $smlnj/ml-yacc/ml-yacc-lib.cm
96    
97    This makes it possible to work with code that requires different
98    versions of these libraries, and which refers to these libraries using
99    their "default" names (i.e., the first set of names above).  In other
100    words, one can un-define or re-define those default names without
101    compromising the proper functioning of the compiler itself.
102    
103    A similar procedure had already been performed for several of the
104    MLRISC libraries that are linked into the compiler.  I did some
105    cleanup on this code.
106    
107    A new file in the config directory (named extrapathconfig) is
108    responsible for setting up path anchors that the compiler itself does
109    not need, but that are typically required by user code.
110    
111    ----------------------------------------------------------------------
112    Name: Matthias Blume
113    Date: 2005/11/07 17:05:00 CST
114    Tag: blume-20051107-slice
115    Description:
116    
117    Fixed erroneous out-of-bounds test in the "update" function of
118    various *ArraySlice modules.  (Thanks to Vesa A. Norrman for pointing
119    out the problem.)
120    
121    Pushed some Basis changes through ML-Lex, CML, and eXene.
122    
123    ----------------------------------------------------------------------
124    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
125    Date: 2005/11/07
126    Description:
127    Fixed a Basis incompatibility: The depreciated function Substring.all
128    was removed (use Substring.full instead).
129    
130    ----------------------------------------------------------------------
131    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
132    Date: 2005/11/05
133    Description:
134    Tweaked interval set API in SML/NJ library; see the CHANGES file for details.
135    
136    ----------------------------------------------------------------------
137    Name: Matthias Blume (http://www.tti-c.org/blume/)
138    Date: 2005/11/02 16:25:00 CST
139    Tag: blume-20051102-hex
140    Description:
141    
142    * Runtime system bootstrap code now accepts hex digits in BOOTLIST
143      in either upper case or lower case format.
144    * Pushed changes to names of Pack<N>{Big,Little} structures through
145      CML and eXene.
146    
147    ----------------------------------------------------------------------
148    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
149    Date: 2005/11/02
150    Description:
151    Fixed a Basis incompatibility: Pack<N>{Big,Little} structures should
152    be called PackWord<N>{Big,Little}.
153    
154    ----------------------------------------------------------------------
155    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
156    Date: 2005/10/28
157    Description:
158    Fixed a minor Basis incompatibility: hex digits should be upper case.
159    
160    ----------------------------------------------------------------------
161    Name: Matthias Blume (blume (at) tti - c (dot) org)
162    Date: 2005/10/25 16:00:00 CDT
163    Tag: Release_110_56
164    Description:
165    
166    New working version (110.56).  NEW BOOTFILES!
167    
168    ----------------------------------------------------------------------
169    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
170    Date: 2005/10/25
171    Description:
172    Added interval sets to utility library (signatures INTERVAL_DOMAIN
173    and INTERVAL_SET, and functor IntervalSetFn).
174    
175    ----------------------------------------------------------------------
176    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
177    Date: 2005/10/14
178    Description:
179    Add Zhong Shao's fix for datatype equality functions.
180    
181    ----------------------------------------------------------------------
182    Name: Allen Leung (leunga (at) reservoir (dot) com)
183    Date: 2005/10/14 12:30:00 EST
184    Tag: leunga-20051014-x86-spill
185    Description:
186    
187    Bug fix a bug found by Carl Hauser.
188    There was a typo in the reload code for FCMP in x86SpillInstr.sml.
189    
190    ----------------------------------------------------------------------
191    Name: Allen Leung (leunga (at) reservoir (dot) com)
192    Date: 2005/10/14 11:15:00 EST
193    Tag: leunga-20051014-x86-asm
194    Description:
195    
196    Removed some debugging code in file x86Asm.sml.
197    The function emit_operand was printing out debugging output.
198    
199    ----------------------------------------------------------------------
200    Name: John Reppy
201    Date: 2005/07/27
202    Tag:
203    Description:
204    
205    Fixed ml-lex to recognize "\r" as representing carriage return.
206    
207    ----------------------------------------------------------------------
208    Name: John Reppy
209    Date: 2005/07/27
210    Tag:
211    Description:
212    
213    Fixed ml-yacc to work on files with non-native end-of-line
214    encodings (e.g., Windows text file on a Unix system).
215    
216    ----------------------------------------------------------------------
217    Name: John Reppy
218    Date: 2005/07/20
219    Tag:
220    Description:
221    
222    Added changes from Dominic Evans (oldmanuk (at) gmail (dot) com)
223    to support HPUX 11.
224    
225    ----------------------------------------------------------------------
226    Name: John Reppy
227    Date: 2005/07/06
228  Tag:  Tag:
229  Description:  Description:
230    
231    Changes to the SML/NJ library.  See smlnj-lib/CHANGES for details.
232    
233    ----------------------------------------------------------------------
234    Name: Matthias Blume (blume (at) tti - c (dot) org)
235    Date: 2005/07/06 00:45:00 CDT
236    Tag: blume-20050706-slice-copy
237    Description:
238    
239    Fixed reversed logic for deciding whether to "copy up" or "copy down"
240    in *-array-slice.sml.
241    
242    ----------------------------------------------------------------------
243    Name: Allen Leung (leunga (at) reservoir (dot) com)
244    Date: 2005/05/31 17:00:00 EST
245    Tag: leunga-20050531-cygwin-fault-2
246    Description:
247    
248    A typo in the cygwin code fixed.
249    
250    ----------------------------------------------------------------------
251    Name: Allen Leung (leunga (at) reservoir (dot) com)
252    Date: 2005/05/31 16:47:00 EST
253    Tag: leunga-20050531-cygwin-fault
254    Description:
255    
256    Updated Cygwin's fault/signal handling to match the Windows version.
257    Updated the export list.
258    
259    ----------------------------------------------------------------------
260    Name: Matthias Blume (blume (at) tti - c (dot) org)
261    Date: 2005/05/18 13:35:00 CDT
262    Tag: Release_110_54
263    Description:
264    
265    New working version (110.54).  NEW BOOTFILES!
266    
267    ----------------------------------------------------------------------
268    Name: Matthias Blume (blume (at) tti - c (dot) org)
269    Date: 2005/05/18 11:58:00 CDT
270    Tag: blume-20050518-installer
271    Description:
272    
273    Added support scripts for Mac OS X PackageMaker and modified
274    config/install.sh so that it supports re-dumping a heap image
275    after customization.
276    
277    ----------------------------------------------------------------------
278    Name: Matthias Blume (blume (at) tti - c (dot) org)
279    Date: 2005/05/18 10:55:00 CDT
280    Tag: blume-20050518-realdiv-noovld
281    Description:
282    
283    Un-overloaded / to work around bug in overloading resolution code.
284    
285    ----------------------------------------------------------------------
286    Name: Matthias Blume (blume (at) tti - c (dot) org)
287    Date: 2005/05/16 23:50:00 CDT
288    Tag: blume-20050516-redump-heap
289    Description:
290    
291    Added mechanism for re-creating a heap file for the interactive system
292    after configuration variables have been changed.
293    
294       CM.redump_heap : string -> unit
295    
296    This is much like SMLofNJ.exportML, but starting from the resulting
297    heap does not return to the caller of CM.redump_heap but
298    restarts the interactive system from scratch.  The original call of
299    CM.redump_heap does not return but ends the interactive session.
300    Thus, CM.redump_heap is a lot like SMLofNJ.exportFn.
301    
302    Internally, redump_heap winds the dynamic execution context back to
303    the point where the original heap image was created and re-executes
304    the heap image generation code in the boot code.
305    
306    ----------------------------------------------------------------------
307    Name: Matthias Blume (blume (at) tti - c (dot) org)
308    Date: 2005/05/09 21:55:00 CDT
309    Tag: blume-20050509-word64
310    Description:
311    
312    Added a hack to the existing hack known as Word64 to make fromString
313    behave correctly.  I am still not sure whether Word64.scan will work
314    as specified with respect to the interaction of radix and prefix.
315    
316    ----------------------------------------------------------------------
317    Name: Allen Leung (leunga (at) reservoir (dot) com)
318    Date: 2005/05/04 11:50:00 EST
319    Tag: leunga-20050504-checkgc
320    Description:
321    
322    Added a gc protocol checking phase.  This phase is enabled with
323    the flag "check-gc".  "debug-check-gc" turns on the verbose mode.
324    
325    ----------------------------------------------------------------------
326    Name: Matthias Blume (blume (at) tti - c (dot) org)
327    Date: 2005/05/04 10:45:00 CDT
328    Tag: blume-20050504-intinf
329    Description:
330    
331    Fixed a bug in the implementation of div and mod for IntInf.
332    Thanks to Neophytos Michael for reporting the problem.
333    
334    ----------------------------------------------------------------------
335    Name: Matthias Blume (blume (at) tti - c (dot) org)
336    Date: 2005/05/04 10:35:00 CDT
337    Tag: blume-20050504-join
338    Description:
339    
340    Added a "join" combinator to the ParserComb module in smlnj-lib.cm.
341    
342    ----------------------------------------------------------------------
343    Name: Matthias Blume (blume (at) tti - c (dot) org)
344    Date: 2005/02/28 23:40:00 CST
345    Tag: blume-20050228-mVar
346    Description:
347    
348    Fixed serious bug (brown paper bag variety) in new implementation of
349    structure Atom in CML.  (I had accidentally used a mailbox instead of
350    an mvar, leaving the door open for races.)
351    
352    ----------------------------------------------------------------------
353    Name: Matthias Blume (blume (at) tti - c (dot) org)
354    Date: 2005/02/25 15:00:00 CST
355    Tag: Release_110_53
356    Description:
357    
358    New working version (110.53).  NEW BOOTFILES!
359    
360    ----------------------------------------------------------------------
361    Name: Matthias Blume (blume (at) tti - c (dot) org)
362    Date: 2005/02/25 14:50:00 CST
363    Tag: blume-20050225-susp
364    Description:
365    
366    Brought back SMLofNJ.Susp.  The underlying suspension type is the one
367    implemented in Core -- which means that it is the same as the one used
368    by the lazy extension.
369    
370    ----------------------------------------------------------------------
371    Name: Matthias Blume (blume (at) tti - c (dot) org)
372    Date: 2005/02/24 16:50:00 CST
373    Tag: blume-20050224-cml-atom
374    Description:
375    
376    Simpler and at the same time more general implementation of structure
377    Atom in CML.
378    
379    ----------------------------------------------------------------------
380    Name: Matthias Blume (blume (at) tti - c (dot) org)
381    Date: 2005/02/15 17:35:00 CST
382    Tag: blume-20050215-tools
383    Description:
384    
385    Created new "tools" directory under "src" and moved "TraceDebugProf"
386    there.
387    
388    ----------------------------------------------------------------------
389    Name: Matthias Blume (blume (at) tti - c (dot) org)
390    Date: 2005/02/10 17:55:00 CST
391    Tag: blume-20050210-longlong
392    Description:
393    
394    Implemented "long long" arguments and results for NLFFI.  (Only the
395    PPC/MacOS implementation is complete, the other backends still need to
396    be updated.)
397    
398    ----------------------------------------------------------------------
399    Name: Matthias Blume (blume (at) tti - c (dot) org)
400    Date: 2005/01/24 17:40:00 CST
401    Tag: blume-20050124-mlyacc
402    Description:
403    
404    Minor cleanup in ML-Yacc rule printing mechanism.  This should fix a
405    problem with certain "as" patterns which previously got rendered
406    using incorrect syntax.
407    
408    ----------------------------------------------------------------------
409    Name: Matthias Blume (blume (at) tti - c (dot) org)
410    Date: 2005/01/18 12:00:00 CST
411    Tag: blume-20050118-profile
412    Description:
413    
414    Made time profiling code (interrupt handler) in runtime system aware
415    of new array representation.
416    
417    ----------------------------------------------------------------------
418    Name: Matthias Blume (blume (at) tti - c (dot) org)
419    Date: 2005/01/14 18:00:00 CST
420    Tag: blume-20050114-heap2exec
421    Description:
422    
423    Implemented new (but still experimental) heap2exec facility.  This is
424    tested under Mac OS X and should work under Linux (will test shortly).
425    It will probably also work on the Sparc (will test some time later).
426    
427      - removed old "HACKED_STANDALONE" hack from runtime
428    
429    To be able to test this, uncomment the request for "heap2asm" in
430    config/targets prior to installation.  (Notice that this is different
431    from "heap2exec" mentioned below.  Not a typo.)
432    
433    To perform an actual test, run the command
434    
435       $ bin/heap2exec heapfile execfile
436    
437    (You can put heap2exec on your shell's path.)
438    
439    For example, run
440    
441      $ bin/heap2exec bin/.heap/ml-yacc.ppc-darwin mly
442    
443    This will create a standalone executable called "mly" which you can
444    then invoke directly as a command.
445    
446    ----------------------------------------------------------------------
447    Name: Matthias Blume (blume (at) tti - c (dot) org)
448    Date: 2005/01/07 11:44:00 CST
449    Tag: blume-20050107-mlstring
450    Description:
451    
452    fixed off-by-one error in ML_STRING macro (globals.c)
453    
454    ----------------------------------------------------------------------
455    Name: Matthias Blume (blume (at) tti - c (dot) org)
456    Date: 2004/12/23 18:00:00 CST
457    Tag: blume-20041223-santa
458    Description:
459    
460    Made ml-build script "smarter" (but only very little).
461    
462    ----------------------------------------------------------------------
463    Name: Matthias Blume (blume (at) tti - c (dot) org)
464    Date: 2004/12/21 15:05:00 CST
465    Tag: blume-20041221-longlong
466    Description:
467    
468    * Implemented access to signed and unsigned long long data in NLFFI.
469    (The parameter-passing part of the picture has not complete. But
470    data structure access seems to work.)
471    
472    * Fixed CM's incorrect assumption that the PPC is little-endian.
473      (On the Mac, it is big-endian.  And that's currently our only
474      PPC platform.)
475    
476    ----------------------------------------------------------------------
477    Name: Matthias Blume (blume (at) tti - c (dot) org)
478    Date: 2004/12/21 12:50:00 CST
479    Tag: blume-20041221-memory
480    Description:
481    
482    Some cleanup in the $c/memory.cm library: separated some concerns by
483    moving allocation code and memory access code each into their own
484    files.
485    
486    ----------------------------------------------------------------------
487    Name: Allen Leung (leunga (at) reservoir (dot) com)
488    Date: 2004/12/17 16:12:00 EST
489    Tag: leunga-20041217-cygwin-smlnj-home
490    Description:
491    
492        The Unix I/O library of SML/NJ on cygwin does not understand
493    Windows style pathname, so problems arise when SMLNJ_HOME is set
494    to a Windows style pathname.  _run-sml now converts SMLNJ_HOME
495    to a POSIX pathname on cygwin.
496    
497    ----------------------------------------------------------------------
498    Name: Matthias Blume (blume (at) tti - c (dot) org)
499    Date: 2004/12/16 13:35:00 CST
500    Tag: Release_110_52
501    Description:
502    
503    Last-minute changes incorporated into 110.52.  Release tag moved.
504    
505    The changes:
506       - HashString.hashString' -> HashString.hashSubstring
507       - bug fix in UnivariateStats
508    
509    ----------------------------------------------------------------------
510    Name: Matthias Blume (blume (at) tti - c (dot) org)
511    Date: 2004/12/15 23:40:00 CST
512    Tag: blume-20041215-hashSubstring
513    Description:
514    
515    - HashString.hashString' -> HashString.hashSubstring
516    - corresponding changes in atom.sml
517    - "de-compressed" (aka. un-obfuscated) code for UnivariateStats and
518      added some comments
519    
520    ----------------------------------------------------------------------
521    Name: Matthias Blume (blume (at) tti - c (dot) org)
522    Date: 2004/12/15 15:30:00 CST
523    Tag: (Release_110_52)
524    Description:
525    
526    New working version (110.52).  NEW BOOTFILES!
527    
528    ----------------------------------------------------------------------
529    Name: Matthias Blume (blume (at) tti - c (dot) org)
530    Date: 2004/12/15 12:45:00 CST
531    Tag: blume-20041215-spaces
532    Description:
533    
534    More on the space problem (this time for Win32).
535    
536    ----------------------------------------------------------------------
537    Name: Matthias Blume (blume (at) tti - c (dot) org)
538    Date: 2004/12/14 17:30:00 CST
539    Tag: blume-20041214-spaces
540    Description:
541    
542    Hacked some of the scripts (in particular: the installer) to cope with
543    spaces in filenames a bit better.  But beware: the current "solution"
544    is likely still full of bugs and inherently incomplete.  (We need to
545    do away with those shell scripts for a comprehensive solution.)
546    
547    ----------------------------------------------------------------------
548    Name: Matthias Blume (blume (at) tti - c (dot) org)
549    Date: 2004/12/13 14:45:00 CST
550    Tag: blume-20041213-ml-makedepend
551    Description:
552    
553    Fixed bug in code for ml-makedepend.
554    
555    ----------------------------------------------------------------------
556    Name: Matthias Blume (blume (at) tti - c (dot) org)
557    Date: 2004/12/09 16:30:00 CST
558    Tag: blume-20041209-statistics
559    Description:
560    
561    Added two simple but potentially useful statistics modules to SML/NJ Library.
562    (See CHANGES file there.)
563    
564    ----------------------------------------------------------------------
565    Name: Matthias Blume (blume (at) tti - c (dot) org)
566    Date: 2004/12/01 16:50:00 CST
567    Tag: blume-20041201-atom
568    Description:
569    
570    smlnj-lib:
571    
572      Added function HashString.hashString' for substrings.
573      Hand-inlined CharVector.foldl into HashString (for speed).
574      Modified implementation of structure Atom to avoid extracting
575      strings from substrings unless necessary.
576      (Also see CHANGES file for smlnj-lib.)
577    
578    ----------------------------------------------------------------------
579    Name: Matthias Blume (blume (at) tti - c (dot) org)
580    Date: 2004/11/24 22:15:00 CST
581    Tag: blume-20041124-cml
582    Description:
583    
584    Made sure CML compiles when Position = Int64.
585    
586    ----------------------------------------------------------------------
587    Name: Matthias Blume (blume (at) tti - c (dot) org)
588    Date: 2004/11/24 14:30:00 CST
589    Tag: blume-20041124-position
590    Description:
591    
592    The compiler can now be compiled in a mode that makes structure Position
593    equal to Int64.  The default, however, is unchanged (Position = Int31)
594    for the time being.
595    
596    To enable 64-bit positions, use the following procedure:
597    
598    1. Start sml
599    2. Autoload $smlnj/cmb.cm (if not already autoloaded)
600    3. Type
601          #set (CMB.symval "USE_64_BIT_POSITIONS") (SOME 1);
602    4. Run CMB.make() as usual.
603    
604    This is barely tested.  The only test so far was a little SML program
605    counting the number of characters in an 8-gigabyte file by
606    reading it character-by-character.  That test was successful.
607    
608    In support of 64-bit positions, a number of new functions have been
609    added to the runtime system.
610    
611    ----------------------------------------------------------------------
612    Name: Matthias Blume (blume (at) tti - c (dot) org)
613    Date: 2004/11/23 14:45:00 CST
614    Tag: blume-20041123-useFile
615    Description:
616    
617    Fixed a problem with unhelpful error messages related to problems with
618    .cm- or .sml files that appear as part of the sml command line.
619    
620    ----------------------------------------------------------------------
621    Name: Matthias Blume (blume (at) tti - c (dot) org)
622    Date: 2004/11/18 15:40:00 CST
623    Tag: Release_110_51
624    Description:
625    
626    New working version (110.51).  NEW BOOTFILES!
627    
628    ----------------------------------------------------------------------
629    Name: Matthias Blume (blume (at) tti - c (dot) org)
630    Date: 2004/11/18 15:35:00 CST
631    Tag: <none>
632    Description:
633    
634    Enabled dlopen and friends for FreeBSD (as recommended by Johannes 5
635    Joemann).
636    
637    ----------------------------------------------------------------------
638    Name: Allen Leung (leunga (at) reservoir (dot) com)
639    Date: 2004/11/17 16:05:21 EST 2004
640    Tag: leunga-20041117-mlrisc-live-kill
641    Description:
642    
643       Added support for MLTree constructs LIVE and KILL
644    to all the architectures.
645    
646    ----------------------------------------------------------------------
647    Name: Matthias Blume (blume (at) tti - c (dot) org)
648    Date: 2004/11/13 00:20:00 CST
649    Tag: blume-20041113-versiontool
650    Description:
651    
652    - Stripped down the versiontool: It now only handles the version number.
653      The date string is generated at bootstrap time (during makeml).
654    
655    - In a previous commit, fixed a minor issue with how polyequal is being
656      translated.  In particular, the code now "looks through" abstractions.
657      This results in slightly fewer polyEqual warnings and hopefully slightly
658      more efficient code.  Important examples for where this matters are
659      the new int64 and word64 types.
660    
661    ----------------------------------------------------------------------
662    Name: Matthias Blume (blume (at) tti - c (dot) org)
663    Date: 2004/11/12 00:30:00 CST
664    Tag: blume-20041112-int64
665    Description:
666    
667    Structure Int64 fully hooked in.  (The implementation is not very
668    efficient, though.)
669    
670    ----------------------------------------------------------------------
671    Name: Matthias Blume (blume (at) tti - c (dot) org)
672    Date: 2004/11/11 17:30:00 CST
673    Tag: blume-20041111-more64
674    Description:
675    
676    All the pieces of Word64 are now there, with the exception of the
677    conversions from and to LargeWord. (Eventually these need to be identities,
678    but for the time being they don't even make sense because LargeWord is
679    32-bit wide.)
680    
681    Also started to add similar support for Int64, but major pieces of that
682    are still missing.
683    
684    ----------------------------------------------------------------------
685    Name: Matthias Blume (blume (at) tti - c (dot) org)
686    Date: 2004/11/11 00:15:00 CST
687    Tag: blume-20041111-word64
688    Description:
689    
690    Structure Word64 is now (almost) complete, word literals and patterns
691    seem to work.  There are a few odd pieces missing.  In particular,
692    I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
693    at the moment.
694    
695    Making Word64 official would mean that LargeWord becomes Word64.  But
696    this requires extreme care because most word-word conversions have to
697    go through LargeWord, so making a mistake means loss of efficiency or
698    worse.  Eventually there will be a solution similar to (but actually
699    simpler than) what I did with IntInf.
700    
701    ----------------------------------------------------------------------
702    Name: Matthias Blume (blume (at) tti - c (dot) org)
703    Date: 2004/11/10 18:12:00 CST
704    Tag: blume-20041110-64bit
705    Description:
706    
707    More 64-bit hacking (but still not even half-way there yet).
708    Also, some assorted improvements to the handling of 8-bit words.
709    
710    ----------------------------------------------------------------------
711    Name: Matthias Blume (blume (at) tti - c (dot) org)
712    Date: 2004/11/09 17:50:00 CST
713    Tag: <none>
714    Description:
715    
716    Started adding some infrastructure for supporting 64-bit int- and
717    word-types.  (Still in its very early stages.)
718    
719    ----------------------------------------------------------------------
720    Name: Matthias Blume (blume (at) tti - c (dot) org)
721    Date: 2004/10/28 10:45:00 CDT
722    Tag: Release_110_50
723    Description:
724    
725    New working version (110.50).  NEW BOOTFILES!
726    
727                 =====================
728    
729    Also:
730    
731      - Changend config/srcarchiveurl from a file just
732        containing the URL string into a file containing
733        shell script code.  The code has access to the $VERSION variable.
734      - Made corresponding changes to config/install.sh and config/unpack.
735      - Default contents of config/srcarchiveurl uses $VERSION and
736        normally does not have to be edited to reflect a version change.
737    
738      (As a result, a version change can be done by just editing
739       config/version, the rest is now automatic.)
740    
741    ----------------------------------------------------------------------
742    Name: Matthias Blume (blume (at) tti - c (dot) org)
743    Date: 2004/10/27 17:50:00 CDT
744    Tag: blume-20041027-btrace-msg
745    Description:
746    
747    BackTrace.monitor now also reports the source of the exception that
748    triggered the trace.
749    
750    ----------------------------------------------------------------------
751    Name: Matthias Blume (blume (at) tti - c (dot) org)
752    Date: 2004/10/27 17:20:00 CDT
753    Tag: blume-20041027-x86-c-calls
754    Description:
755    
756    This is the HISTORY entry for two earlier commits, both concerning
757    the x86 c-calls code in MLRISC:
758    
759      - added a missing LOAD in the code that deals with struct arguments
760      - made sure the caller does not add the wrong number of bytes to the
761        stack pointer after a call of a function returning a struct
762        (the callee already pops the implicit argument which points to
763         the space reserved for the result)
764    
765    ----------------------------------------------------------------------
766    Name: Allen Leung (leunga (at) reservoir (dot) com)
767    Date: 2004/10/24 14:00:00 EST
768    Tag: leunga-20041024-x86-gas-fucomip
769    Description:
770    
771       John discovered a bug in the syntax of fucomip.
772       The opcodes FU?COMIP? have been changed to
773    
774           fu?comip? %st(i), %st
775    
776    ----------------------------------------------------------------------
777    Name: Matthias Blume (blume (at) tti - c (dot) org)
778    Date: 2004/10/20 15:06:00 CDT
779    Tag: blume-20041020-standalone-backtrace
780    Description:
781    
782    Added a mechanism for getting back-trace information from standalone
783    programs.  Here is how it works:
784    
785       1. The part of the program from which you want to get backtrace
786          information (usually the whole program) should be wrapped with
787          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
788          main program could be modified from something like
789    
790             fun main (pgm, args) = ...
791    
792         to
793    
794            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
795    
796       2. To be able to access BackTrace.monitor, you have to add
797          library $smlnj-tdp/plugins.cm to the .cm file that contains your
798          main function.
799    
800       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
801          CM might have created in the past for your project).
802    
803       4. Build the system using this command line:
804    
805             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
806                      myprog.cm MyProg.main myprog
807    
808          instead of the usual
809    
810             ml-build myprog.cm MyProg.main myprog
811    
812    I changed a library name:
813    
814       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
815    
816    New libraries:
817    
818       $smlnj-tdp/back-trace.cm
819          -- when loaded causes the back-trace plugin to be installed
820       $smlnj-tdp/coverage.cm
821          -- when loaded causes the coverage plugin to be installed
822    
823    ----------------------------------------------------------------------
824    Name: Matthias Blume (blume (at) tti - c (dot) org)
825    Date: 2004/10/18 16:45:00 CDT
826    Tag: blume-20041018-groupowner
827    Description:
828    
829    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
830    
831    Eliminated group owner specs from .cm files throughout the source tree.
832    
833    ----------------------------------------------------------------------
834    Name: Matthias Blume (blume (at) tti - c (dot) org)
835    Date: 2004/10/15 15:45:00 CDT
836    Tag: blume-20041015-coverage
837    Description:
838    
839    * Test coverage tool added!
840    * Further reorganization of tracing-, debugging-, and profiling support:
841    
842        - moved original BTImp -- now called BackTrace -- into a separate
843          library called $/trace-debug-profile.cm
844        - eliminated all mentions of BTrace from SMLofNJ.Internals
845        - only the instrumentation mechanism is now left in the compiler proper
846        - BackTrace module is a plugin which is NOT plugged in by default
847        - Coverage module is another such plugin
848    
849    To get the benefits of any of these plugin modules, the code in
850    question must be compiled with tdp instrumentation turned on.  This
851    can be done by setting SMLofNJ.Internals.TDP.mode to true.
852    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
853    
854    Plugins are selected at link time.  (Pre-compiled instrumented code
855    can be re-loaded with different plugins in effect.)  When an
856    instrumented module is linked, whatever plugins are at that time
857    enabled will come into effect for that module.
858    
859    To enable the back-trace plugin, load library $/trace-debug-profile.cm
860    and invoke BackTrace.install() (e.g., from the interactive prompt).
861    To enable the coverage plugin, load the same library and invoke
862    Coverage.install().
863    
864    Back-traces are generated automatically on uncaught exceptions and
865    when the code in question explicitly invokes BackTrace.trigger().
866    
867    Coverage (and execution frequency-) information must be queried
868    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
869    
870    ----------------------------------------------------------------------
871    Name: Matthias Blume (blume (at) tti - c (dot) org)
872    Date: 2004/10/14 17:40:00 CDT
873    Tag: blume-20041014-tdp-core
874    Description:
875    
876    Snapshot of a significant overhaul of how the trace/debug/profile support
877    is hooked into the system (specifically: Core and SMLofNJ.Internals).
878    
879    ----------------------------------------------------------------------
880    Name: Matthias Blume (blume (at) tti - c (dot) org)
881    Date: 2004/10/13 16:34:00 CDT
882    Tag: blume-20041013-tdp
883    Description:
884    
885    Some rationalization of names:
886    
887       structure BTrace -> structure TDPInstrument
888       etc.
889    
890    This is is preparation of using the original back-trace
891    instrumentation for other purposes.  "TDP" stands for
892    Trace/Debug/Profile.
893    
894    The control flag controlling whether instrumentation is on or off is now
895    registered under a different name, so instead of running sml as
896    
897     sml -Cinstrument.btrace-mode=true
898    
899    one has to say
900    
901     sml -Ctdp.instrument=true
902    
903    ----------------------------------------------------------------------
904    Name: Matthias Blume (blume (at) tti - c (dot) org)
905    Date: 2004/10/11 16:37:00 CDT
906    Tag: blume-20041011-regions
907    Description:
908    
909    Made some minor modifications to elabcore.sml to have source regions
910    be propagated more tightly -- resulting in better (i.e., smaller)
911    regions being reported in error- and debug messages.
912    
913    ----------------------------------------------------------------------
914    Name: Matthias Blume (blume (at) tti - c (dot) org)
915    Date: 2004/10/08 22:50:00 CDT
916    Tag: blume-20041008-cmkw
917    Description:
918    
919    Fixed handling of keywords in .cm files: After seeing "is" the lexer
920    treats subsequent occurrences of "group", "library", "source", "is",
921    "*", and "-" as ordinary identifiers rather than keywords.
922    
923    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
924    is supposed to accept a tool argument called "source", but this did
925    not work because of the clash with the keyword.
926    
927    ----------------------------------------------------------------------
928    Name: Matthias Blume (blume (at) tti - c (dot) org)
929    Date: 2004/10/07 16:00:00 CDT
930    Tag: blume-20041007-cleanup
931    Description:
932    
933    Assorted cleanup work:
934    
935      - got rid of intstrmap in favor of using the library's
936        hash table implementation
937      - threw out most of the pathnames stuff, as it was not used anyway
938      - simplified tokentable implementation
939      - fixed some minor spelling errors
940    
941    ----------------------------------------------------------------------
942    Name: Matthias Blume (blume (at) tti - c (dot) org)
943    Date: 2004/10/06 15:15:15 CDT
944    Tag: blume-20041006-handler
945    Description:
946    
947    Cleaned up the absyn to reflect the invariant that HANDLE always
948    carries a FNexp as part of the type definition.  This eliminates some
949    superfluous sanity checks at runtime down the road.
950    
951    Some minor cleanup of the btrace code.
952    
953    ----------------------------------------------------------------------
954    Name: Matthias Blume (blume (at) tti - c (dot) org)
955    Date: 2004/10/01 10:20:30 CDT
956    Tag: blume-20041001-slave
957    Description:
958    
959    Added hack to make slave mode work in the presence of the version
960    tool.  (Still, since the master does two passes over the code for
961    CMB.make, the release number gets bumped twice when slaves are
962    attached.  I don't know if this is worth fixing...)
963    
964    ----------------------------------------------------------------------
965    Name: Matthias Blume (blume (at) tti - c (dot) org)
966    Date: 2004/09/30 10:55:00 CDT
967    Tag: blume-20040930-version
968    Description:
969    
970    * Moved the "version" magic into its own little library under
971      src/system/smlnj/internal.  This avoids expensive reconstruction of
972      a stable src/compiler/core.cm.
973    
974    * At the same time, structure CompilerVersion is now known as
975      structure SMLNJVersion.
976    
977    * Arranged for the version tool to NOT kick in when rebuilding the system
978      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
979      Also, loading the versiontool does not work when rebuilding the system
980      because CM is not properly initialized at that time.
981    
982    ----------------------------------------------------------------------
983    Name: Matthias Blume (blume (at) tti - c (dot) org)
984    Date: 2004/09/29 14:00:00 CDT
985    Tag: blume-20040929-autoversion
986    Description:
987    
988    Implemented some CM magic to have
989        file src/compiler/TopLevel/main/version.sml
990    generated automagically.
991    The version is taken from two files: config/version and config/release.
992    The first is expected to contain a two-part version number such as 110.49.
993    The second should contain a single number, but it may be missing.
994    
995    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
996    time the version tool is loaded (which is the first time you say CMB.make),
997    then the tool will increment the value stored in config/release every
998    time CMB.make is invoked.
999    
1000    The binfile format is now insensitive to anything beyond the first
1001    two components of a version number, so bumping the release does not render
1002    binfiles incompatible.  Auto-bumping can be used to keep track of versions
1003    during development without invalidating existing binfiles.
1004    
1005    In any case, every CMB.make updates the date information in version.sml.
1006    (This is the date that is printed in the banner.)
1007    
1008    ----------------------------------------------------------------------
1009    Name: Matthias Blume (blume (at) tti - c (dot) org)
1010    Date: 2004/09/28 10:53:00 CDT
1011    Tag: blume-20040928-controls
1012    Description:
1013    
1014    Some cleanup of the controls code.
1015    
1016    ----------------------------------------------------------------------
1017    Name: Matthias Blume (blume (at) tti - c (dot) org)
1018    Date: 2004/09/27 22:08:00 CDT
1019    Tag: blume-20040927-controls
1020    Description:
1021    
1022    Added two pieces of functionality to the Controls interface:
1023    
1024    1.   val save'restore: 'a control -> unit -> unit
1025    
1026       grabs the current value of the control in stage 1 and restores it
1027       in stage 2.
1028    
1029    2.   val set' : 'a control * 'a -> unit -> unit
1030    
1031       stores the given value into the control in stage 2 (i.e., delayed)
1032       but does all error checking in stage 1.
1033       (This is for string controls that need to do parse their argument --
1034        something that might fail.  In some cases, notably in CM, one
1035        already knows the intended argument but wants to delay the actual
1036        assignment until a time when error recovery would be more difficult.)
1037    
1038    Changed the handling of controls in tool arguments to classes "sml" and
1039    "lazysml":
1040       - use Controls.save'restore as a more robust way of restoring the
1041         old value (in particular: without having to re-parse the string)
1042       - use controls to handle the "overload" keyword in the init group
1043         (I believe this change actually fixes a long-standing obscure bug.)
1044    
1045    ----------------------------------------------------------------------
1046    Name: Matthias Blume (blume (at) tti - c (dot) org)
1047    Date: 2004/09/27 17:00:00 CDT
1048    Tag: blume-20040927-lazysml
1049    Description:
1050    
1051    Added a new tool class called "lazysml" to CM's tool chest.  The only
1052    difference to "sml" is that compilation is done with Control.lazysml
1053    set to true.  A source of class "lazysml" is automatically recognized
1054    by a file name suffix of ".lml".
1055    
1056    In addition to the above feature, the original class "sml" now also
1057    supports a tool argument "lazy" which has the same effect.  As a
1058    result, the following three lines are equivalent:
1059    
1060        foo.sml : lazysml
1061        foo.sml : sml (lazy)
1062        foo.sml (lazy)
1063    
1064    The setting goes into effect both during parsing and during
1065    compilation.  The original setting is restored right after parsing and
1066    after compilation, respectively.
1067    
1068    In addition to all the above, there is also a general mechanism to set
1069    ANY of the "controls" that are available at the command line via
1070    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
1071    apply as well. (In fact, "lazy" is implemented as a special case of
1072    the general mechanism.)
1073    
1074    The .cm file syntax uses a new keyword tool argument called "with".
1075    There are several ways of indicating the desired settings:
1076    
1077        foo.sml (with:parser.quotations=true)
1078        foo.sml (with:(name:parser.quotations value:true))
1079        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
1080        foo.sml (with:(name1=value1 name2=value2 ...))
1081        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
1082    
1083        etc.
1084    
1085    Another possible abbreviation is to leave out the =v or value:v part
1086    if the name refers to a boolean control (in which case the value is
1087    taken to be true).  Thus, one could get lazy sml also by saying:
1088    
1089        foo.sml (with:parser.lazy-keyword=true)
1090        foo.sml (with:parser.lazy-keyword)
1091        foo.sml (with:(name:parser.lazy-keyword value:true))
1092        foo.sml (with:(name:parser.lazy-keyword))
1093    
1094    ----------------------------------------------------------------------
1095    Name: Matthias Blume (blume (at) tti - c (dot) org)
1096    Date: 2004/09/24 16:22:00 CDT
1097    Tag: blume-20040924-ppc-long-branch
1098    Description:
1099    
1100    Turned message about "emiting long form of branch" off by default.
1101    Added a control flag to turn it back on when desired.
1102    
1103    ----------------------------------------------------------------------
1104    Name: Matthias Blume (blume (at) tti - c (dot) org)
1105    Date: 2004/09/24 16:05:00 CDT
1106    Tag: blume-20040924-rounding
1107    Description:
1108    
1109    Applied patch for setting rounding modes under Mac OS X.  Thanks to
1110    Melissa O'Neill for providing the code!
1111    
1112    ----------------------------------------------------------------------
1113    Name: Matthias Blume (blume (at) tti - c (dot) org)
1114    Date: 2004/09/23 17:30:00 CDT
1115    Tag: blume-20040923-envvars
1116    Description:
1117    
1118    1. Changed definition of type ControlRegistry.registry_tree to
1119       include control_info (i.e., the name of the controlling environment
1120       variable).
1121    
1122    2. Added command-line flags -e and -E to print the names of environment
1123       variables that can be used to control internal settings.
1124       (This uses the new API mentioned in 1.)
1125    
1126    ----------------------------------------------------------------------
1127    Name: Matthias Blume (blume (at) tti - c (dot) org)
1128    Date: 2004/09/13 16:50:00 CDT
1129    Tag: Release_110_49
1130    Description:
1131    
1132    New working version (110.49).  NEW BOOTFILES!
1133    
1134    ----------------------------------------------------------------------
1135    Name: Matthias Blume
1136    Date: 2004/09/13 16:20:00 CDT
1137    Tag: blume-20040913-config-mlrisc
1138    Description:
1139    
1140    Put target "mlrisc" back into the default list.
1141    (There is no harm in having it, and some users have expressed their
1142    wish to have "mlrisc" included by default.)
1143    
1144    ----------------------------------------------------------------------
1145    Name: John Reppy
1146    Date: 2004/09/13
1147    Tag: jhr-20040913-signals
1148    Description:
1149    
1150    Fixed the signal masking code to properly nest mask/unmask operations
1151    on a per-signal basis.
1152    
1153    ----------------------------------------------------------------------
1154    Name: Matthias Blume (blume (at) tti - c (dot) org)
1155    Date: 2004/09/08 13:20:00 CDT
1156    Tag: blume-20040908-heap-magic
1157    Description:
1158    
1159    Bumped the heap macig number to 0x09082004 to account for the changed
1160    layout of the ML frame under MacOS X.
1161    
1162    ----------------------------------------------------------------------
1163    Name: Allen Leung (leunga (at) reservoir (dot) com)
1164    Date: 2004/09/03 11:26:00 EST
1165    Tag: leunga-20040903-cygwin-install
1166    Description:
1167    
1168       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
1169    The Cygwin runtime is turned on by setting the environment
1170    variable SMLNJ_CYGWIN_RUNTIME to 1.
1171    
1172    ----------------------------------------------------------------------
1173    Name: Matthias Blume (blume (at) tti - c (dot) org)
1174    Date: 2004/08/31 17:15:00 CDT
1175    Tag: blume-20040831-core
1176    Description:
1177    
1178    Added some exports to src/compiler/core.cm upon request by J. Joemann.
1179    
1180    ----------------------------------------------------------------------
1181    Name: Matthias Blume (blume (at) tti - c (dot) org)
1182    Date: 2004/08/30 17:55:00 CDT
1183    Tag: blume-20040830-installer
1184    Description:
1185    
1186    Upon request by Johannes Joemann:
1187    
1188    - improved ML code of installer to fall back to coping when renaming
1189      fails (i.e., when source and target are on different file systems);
1190      the code compiles but has yet to be tested in anger
1191    - removed mlrisc from list of default targets (config/targets)
1192    
1193    ----------------------------------------------------------------------
1194    Name: Matthias Blume (blume (at) tti - c (dot) org)
1195    Date: 2004/08/27 17:20:00 CDT
1196    Tag: blume-20040827-ptreql
1197    Description:
1198    
1199    Added ptreql primop to structure InlineT (upon request from Larry
1200    Paulson).
1201    
1202    ----------------------------------------------------------------------
1203    Name: Allen Leung (leunga (at) reservoir (dot) com)
1204    Date: 2004/08/15 21:21:00 EST
1205    Tag: leunga-110_48-udgraph
1206    Description:
1207    
1208    Another bug fix from Carl Hauser:
1209    
1210    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
1211    > 48c48
1212    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
1213    > e::L)
1214    > ---
1215    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
1216    > Without this, any deletion of an edge in an undirected graph does severe
1217    > violence to the graph.
1218    
1219    ----------------------------------------------------------------------
1220    Name: Allen Leung (leunga (at) reservoir (dot) com)
1221    Date: 2004/08/10 23:35:00 EST
1222    Tag: leunga-110_48-ppc
1223    Description:
1224    
1225       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
1226    
1227    ----------------------------------------------------------------------
1228    Name: Matthias Blume (blume (at) tti - c (dot) org)
1229    Date: 2004/08/10 12:00:00 CDT
1230    Tag: Release_110_48
1231    Description:
1232    
1233    New working version (110.48).  NEW BOOTFILES!
1234    
1235    ----------------------------------------------------------------------
1236    Name: Allen Leung (leunga (at) reservoir (dot) com)
1237    Date: 2004/08/09 12:21:00 EST
1238    Tag: leunga-110_47-dijsktra
1239    Description:
1240    
1241    Bug fix from Carl Hauser:
1242    
1243    single_source_shortest_paths in dijkstra.sml was observed to get wrong
1244    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
1245    
1246    The problem is that following the expression A.update(dist,s,Num.zero)
1247    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
1248    
1249    ----------------------------------------------------------------------
1250    Name: Matthias Blume (blume (at) tti - c (dot) org)
1251    Date: 2004/08/06 18:10:00 CDT
1252    Tag: blume-20040806-cmdline
1253    Description:
1254    
1255    Fiddled with handling of command-line options:
1256    
1257      * sml now quits after processing the command line
1258        if -H, -S, -h<n>, or -s<n> appears as the last
1259        command-line argument
1260      * a new option -q terminates the session when encountered on
1261        the command line; subsequent arguments will be ignored
1262      * bug fixes: short (erroneous) arguments are no longer ignored
1263        completely
1264    
1265    ----------------------------------------------------------------------
1266    Name: Allen Leung (leunga (at) reservoir (dot) com)
1267    Date: 2004/08/04 18:17:00 EST
1268    Tag: leunga-110_47-ppc-ibm-asm
1269    Description:
1270    
1271      - Added minimal IBM assembly syntax support for PowerPC.
1272    
1273      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
1274        been altered in the runtime.  We need an automatic way to keep the file
1275        in sync.
1276    
1277    ----------------------------------------------------------------------
1278    Name: Matthias Blume (blume (at) tti - c (dot) org)
1279    Date: 2004/08/04 14:00:00 CDT
1280    Tag: Release_110_47
1281    Description:
1282    
1283    New working version (110.47).  NEW BOOTFILES!
1284    
1285    ----------------------------------------------------------------------
1286    Name: Matthias Blume (blume (at) tti - c (dot) org)
1287    Date: 2004/08/03 14:25:00 CDT
1288    Tag: blume-20040803-callingconv
1289    Description:
1290    
1291    Added low-level support for choosing C calling conventions by
1292    twiddling the type of rawccall.  (See
1293    src/compiler/Semant/types/cproto.sml for details.)
1294    
1295    ----------------------------------------------------------------------
1296    Name: Matthias Blume (blume (at) tti - c (dot) org)
1297    Date: 2004/08/02 15:55:00 CDT
1298    Tag: blume-20040802-backout
1299    Description:
1300    
1301    Backed out of change to win32-filesys.c.  The earlier patch to
1302    get_file_time caused CM to produce files with the wrong time stamp.
1303    
1304    ----------------------------------------------------------------------
1305    Name: Matthias Blume (blume (at) tti - c (dot) org)
1306    Date: 2004/08/02 14:45:00 CDT
1307    Tag: blume-20040802-nlffi-win32
1308    Description:
1309    
1310    Added NLFFI support for Win32, adapted from a patch provided by David
1311    Hansel.  This is currently completely untested.  Also, the issue
1312    concerning stdcall vs. ccall is still unresolved.
1313    
1314    ----------------------------------------------------------------------
1315    Name: Matthias Blume (blume (at) tti - c (dot) org)
1316    Date: 2004/07/30 17:55:00 CDT
1317    Tag: blume-20040730-various
1318    Description:
1319    
1320    Gearing up towards 110.47...
1321    
1322    - various minor bugfixes to ml-nlffigen
1323    - a beginning of a manual for nlffi
1324    
1325    - eliminated 'export name=value' in config/install.sh as this does
1326      not work with certain versions of /bin/sh
1327      (Thanks to David King at Motorola for catching this.)
1328    
1329    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
1330      - added a test for tm==NULL to gmtime.c and localtime.c
1331      - applied patch for incorrect GetFileTime under win32
1332      - toSeconds -> toMilliseconds in Win32/win32-process.sml
1333    
1334    ----------------------------------------------------------------------
1335    Name: Matthias Blume (blume (at) tti - c (dot) org)
1336    Date: 2004/07/21 18:20:00 CDT
1337    Tag: blume-20040721-nlffigen
1338    Description:
1339    
1340    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
1341      for a typedef to an incomplete type, but leave out the "typ" member.
1342      (This is just for consistency.)
1343    - Started to produce what is supposed to become better (i.e., comprehensive)
1344      documentation of what ml-nlffigen does and produces.
1345    
1346    ----------------------------------------------------------------------
1347    Name: Matthias Blume (blume (at) tti - c (dot) org)
1348    Date: 2004/07/14 16:25:00 CDT
1349    Tag: blume-20040714-union
1350    Description:
1351    
1352    Added C_UNION to c-calls/c-types.sml and updated the machinery
1353    (ml-nlffigen, cproto.sml) that conveys C function interface
1354    information to the code generator.
1355    
1356    However, the actual architecture-specific implementation of function
1357    arguments and results that are C unions is still not implemented.
1358    
1359    ----------------------------------------------------------------------
1360    Name: Allen Leung (leunga (at) reservoir (dot) com)
1361    Date: 2004/07/14 14:38:00 EST
1362    Tag: leunga-110_46_1-ppc-lwzu
1363    Description:
1364    
1365        Added these instructions to the PowerPC architecture:
1366        LBZU(X), LHZU(X), LWZU(X),
1367        STWU(X), STFDU, STFSU
1368    
1369        etc...
1370    
1371        Note: I haven't added their instruction encoding into the description.
1372    
1373    ----------------------------------------------------------------------
1374    Name: Allen Leung (leunga (at) reservoir (dot) com)
1375    Date: 2004/07/13 15:04:00 EST
1376    Tag: leunga-110_46_1-ppc-lwarx
1377    Description:
1378    
1379       Added the two instructions LWARX and STWCX to the PowerPC
1380    instruction set.
1381    
1382       A (untested) rewrite of loop-structure.sml.  The old version
1383    is completely broken.
1384    
1385    ----------------------------------------------------------------------
1386    Name: Matthias Blume (blume (at) tti - c (dot) org)
1387    Date: 2004/07/13 13:50:00 CDT
1388    Tag: blume-20040713-nlffi
1389    Description:
1390    
1391    - use paramAlloc to report c-calls with too many arguments
1392      (for PPC version where parameter area is pre-allocated)
1393    - added ccall_maxargspace to machspec (to implement the above)
1394    - made "make" commend in CM's "make" tool configurable
1395    - added option (default: on) for passing the name of the SML/NJ's "bin"
1396      directory to "make"; the call looks like this:
1397    
1398         make <options> SMLNJ_BINDIR=<dir> <target>
1399    
1400      This can be used by the Makefile to, e.g., pick the "right" version
1401      of ml-nlffigen.
1402    - minor code tweaks
1403    
1404    ----------------------------------------------------------------------
1405    Name: Matthias Blume (blume (at) tti - c (dot) org)
1406    Date: 2004/07/12 22:50:00 CDT
1407    Tag: blume-110_46_1-macosx-nlffi
1408    Description:
1409    
1410    NLFFI under Mac OS X now working (sort of).  This is largely untested,
1411    though.
1412    
1413    Note:  1. You have to make a new, clean build of the runtime system.
1414           2. There are new BOOTFILES, you have to use them!
1415              (Doing the bootstrap process yourself would be *very* painful!
1416               If you absolutely have to do it, build the system under
1417               a different architecture and then cross-compile.)
1418    
1419    Version bumped to 110.46.1 to account for runtime data format changes.
1420    
1421    ----------------------------------------------------------------------
1422    Name: Matthias Blume (blume (at) tti - c (dot) org)
1423    Date: 2004/06/18 14:30:00 CDT
1424    Tag: blume-20040618-unix
1425    Description:
1426    
1427    Changed the implementation of structure Unix so that the same stream
1428    is returned every time one of the {text,bin}{In,Out}streamOf functions
1429    is invoked on the same proc.  This is not what the spec currently
1430    says -- although IMO it arguably should.  (See discussion below.)
1431    
1432    ----------------------------------------------------------------------
1433    Name: Matthias Blume (blume (at) tti - c (dot) org)
1434    Date: 2004/06/17 18:15:00 CDT
1435    Tag: Release_110_46
1436    Description:
1437    
1438    New working version (110.46).  NEW BOOTFILES!
1439    
1440    ----------------------------------------------------------------------
1441    Name: Matthias Blume (blume (at) tti - c (dot) org)
1442    Date: 2004/06/17 17:20:00 CDT
1443    Tag: blume-20040617-timer-unix
1444    Description:
1445    
1446    Changed the interface of structures Timer and Unix to match the most
1447    recent Basis spec.
1448    
1449    In the case of Unix there still seems to be an open/weird issue:
1450    
1451       The {text,bin}{In,Out}streamOf functions are supposed to create
1452       fresh streams whenever they are called -- as opposed to have them
1453       return the same stream every time.  This design is supposed to
1454       prevent space leaks caused by proc values hanging on to streams.
1455    
1456       The reap function, on the other hand, is supposed to close the
1457       streams.  This cannot be done without having a handle on the
1458       stream in proc after all...
1459    
1460    I took the liberty to implement the following stopgap solution:
1461    
1462      The proc value hangs on to the most recently created stream(s).
1463      Reap closes those.  If either or both of the two streams hadn't
1464      been created at all yet, then reap will close the corresponding
1465      file descriptors directly.
1466    
1467    PS: I don't understand the original space leak argument anymore.  If
1468    a proc hangs on to the imperative stream, then I/O operations on those
1469    will advance the state of the cached stream and avoid the space leak.
1470    
1471    ----------------------------------------------------------------------
1472    Name: Matthias Blume (blume (at) tti - c (dot) org)
1473    Date: 2004/05/28 16:45:00 CDT
1474    Tag: blume-20040528-basis
1475    Description:
1476    
1477    Added signature PACK_REAL and exported functor PrimIO.
1478    
1479    ----------------------------------------------------------------------
1480    Name: Matthias Blume (blume (at) tti - c (dot) org)
1481    Date: 2004/05/25 16:00:00 CDT
1482    Tag: blume-20040525-group-owner
1483    Description:
1484    
1485    CM now ignores (but still accepts) the "owner" information in group
1486    descriptions.  The owner of a group is its next enclosing
1487    library. Each group must have a unique owner.  (There is a virtual
1488    "toplevel" library that owns groups which are not nested within a real
1489    library.)  Previously, each group had to explicitly declare its owner,
1490    and CM would check that such a declaration is correct.  The new scheme
1491    is to have CM check that for each group there is precisely one owning
1492    library.
1493    
1494    The advantage of the new scheme is that the programmer no longer needs
1495    to maintain the somewhat annoying owner information.  The downside is
1496    that CM cannot enforce the ownership rule across multiple runs of
1497    CM.make.  Fortunately, enclosing the same group in two different
1498    libraries A and B which are not part of the same program does not
1499    cause real problems.
1500    
1501    ----------------------------------------------------------------------
1502    Name: Matthias Blume (blume (at) tti - c (dot) org)
1503    Date: 2004/05/20 16:00:00 CDT
1504    Tag: blume-20040520-win32
1505    Description:
1506    
1507    Made the win32 version work again.  (Strangely, a misplaced comma had
1508    slipped into win32-process.c which prevented the runtime from being
1509    compiled correctly.)
1510    
1511    Also, included a minor addition to ml-build.bat analogous to what was
1512    done in blume-20040519-ml-build.
1513    
1514    ----------------------------------------------------------------------
1515    Name: Matthias Blume (blume (at) tti - c (dot) org)
1516    Date: 2004/05/19 22:10:00 CDT
1517    Tag: blume-20040519-ml-build
1518    Description:
1519    
1520    Arranged for ml-build to clean up after itself a little bit better.
1521    The script generates a temporary SML source file and compiles it using
1522    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
1523    gets rid of those at the end, so they don't accumulate under .cm.
1524    
1525    This required a minor change to install.sh because the name of the
1526    metadata directory (default: .cm) is actually configurable at
1527    installation time.
1528    
1529    ----------------------------------------------------------------------
1530    Name: Matthias Blume (blume (at) tti - c (dot) org)
1531    Date: 2004/05/18 15:50:00 CDT
1532    Tag: blume-20040518-mkreader
1533    Description:
1534    
1535    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
1536    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
1537    
1538    ----------------------------------------------------------------------
1539    Name: Matthias Blume (blume (at) tti - c (dot) org)
1540    Date: 2004/05/11 14:35:00 CDT
1541    Tag: blume-20040511-win32sock
1542    Description:
1543    
1544    Added previously missing support for many socket-related functions
1545    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
1546    for the voluminous patch!
1547    
1548    (I have not tested this patch under win32 yet.)
1549    
1550    Here is David's e-mail:
1551    
1552    Hi,
1553    
1554    Attached to this email you find a diff against sml/nj 110.45
1555    that will enable socket support under Windows.
1556    
1557    To apply the patch (using unix or cygwin)
1558    1) gunzip runtime.diff.gz
1559    2) "cd" into "src/runtime" in the source tree of a fresh
1560       110.45 installation.
1561    3) patch -p 1 < [your/path/to]runtime.diff
1562    
1563    The code compiles fine but has NOT yet been extensively tested.
1564    I only ran a few tests for basic socket client functionality
1565    (which worked fine).  Especially the functions that use ioctl
1566    are not tested at all and might not work (see below).
1567    
1568    I implemented this since we want to move to a newer version of sml/nj
1569    but need socket support in order to use it.  This is the first time I
1570    even had a look at the sml/nj source,  so please review my changes
1571    before making this part of the distribution!  Here are a few issues
1572    that I think might be better for someone to solve who is more
1573    familiar with the sml/nj source (and socket programming):
1574    
1575    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
1576      exception since I could not figure out what the windows equivalent
1577      of these functions is
1578    
1579    - In sockets-osdep.h there are a some #include statements that are
1580      only used in a few files that include sockets-osdep.h
1581    
1582    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
1583      does not process its return value since I don't know how to
1584      report an error upwards.
1585    
1586    - It would probably be good to have a call to WSACleanup() when
1587      the library is unloaded (if there is such a possibility).
1588      Otherwise I think Windows will take care of this automatically
1589      when the process finishes.
1590    
1591    - I used ioctlsocket() as a replacement for ioctl() but I have
1592      no idea if that is actually the proper replacement on Windows.
1593    
1594    - All these issues are marked in the code by "FIXME" comments.
1595    
1596    We use sml/nj extensively in our products and are quite happy
1597    with it.  I hope this contribution will help you.
1598    
1599    Keep up the good work!
1600    
1601    David
1602    
1603    ----------------------------------------------------------------------
1604    Name: Matthias Blume (blume (at) tti - c (dot) org)
1605    Date: 2004/05/11 14:20:00 CDT
1606    Tag: blume-20040511-installml
1607    Description:
1608    
1609    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
1610    the patch.)
1611    
1612    ----------------------------------------------------------------------
1613    Name: Matthias Blume (blume (at) tti - c (dot) org)
1614    Date: 2004/05/11 14:05:00 CDT
1615    Tag: blume-20040511-nlffi-netbsd
1616    Description:
1617    
1618    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
1619    the patch.)
1620    
1621    ----------------------------------------------------------------------
1622    Name: Matthias Blume (blume (at) tti - c (dot) org)
1623    Date: 2004/05/11 12:05:00 CDT
1624    Tag: blume-20040511-exports
1625    Description:
1626    
1627    As per request by Adam Chlipala <adam@hcoop.net>, extended various
1628    export lists in compiler-related .cm-files.
1629    
1630    ----------------------------------------------------------------------
1631    Name: Matthias Blume (blume (at) tti - c (dot) org)
1632    Date: 2004/05/11 11:35:00 CDT
1633    Tag: blume-20040511-allsource
1634    Description:
1635    
1636    The installer now honors the "src-smlnj" target again, although its meaning
1637    has changed from "all sources required for the compiler" to "all sources
1638    the installer knows about".  In other words, if you enable "src-smlnj"
1639    in the "targets" file, then the installer will pull in sources for
1640    everything.  (Notice that this refers to source code only.  Compiled
1641    code is still only installed for modules that were requested explicitly
1642    or which are required for other modules that were requested explicitly.)
1643    
1644    ----------------------------------------------------------------------
1645    Name: Matthias Blume (blume (at) tti - c (dot) org)
1646    Date: 2004/04/23 17:40:00 CDT
1647    Tag: blume-20040423-ieee-scan
1648    Description:
1649    
1650    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
1651    in the exponent calculation we get INF or ZERO (depending on the mantissa
1652    and the sign of the exponent).
1653    
1654    ----------------------------------------------------------------------
1655    Name: Matthias Blume (blume (at) tti - c (dot) org)
1656    Date: 2004/04/23 10:40:00 CDT
1657    Tag: blume-20040423-ml-build
1658    Description:
1659    
1660    The ml-build script now terminates with a non-0 status when something
1661    goes wrong.
1662    
1663    ----------------------------------------------------------------------
1664    Name: Matthias Blume (blume (at) tti - c (dot) org)
1665    Date: 2004/04/22 16:35:00 CDT
1666    Tag: blume-20040422-Option
1667    Description:
1668    
1669    Made exception Option to be the same as exception Option.Option
1670    (as it should be).
1671    
1672    ----------------------------------------------------------------------
1673    Name: Allen Leung (leunga (at) reservoir (dot) com)
1674    Date: 2004/03/19 14:40:00 EST
1675    Tag: leunga-20040319-cygwin-nlffi
1676    Description:
1677    
1678    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1679    of SML/NJ.  The problem is that
1680    
1681       lib = dlopen(NULL, ...)
1682       f   = dlsym(lib, "malloc");
1683    
1684    does not work on Windows unless we explicitly export symbols
1685    such as 'malloc' during linking.  We fixed this by explicitly
1686    exporting the required symbols with the magic gcc incantation:
1687    
1688        -Wl,--export-all cygwin.def
1689    
1690    where cygwin.def is a file containing all the symbols that
1691    we wish to export.
1692    
1693    I suspect this is a Windows problem and we'll have to
1694    do the same (somehow with windows compilers) when
1695    we build the native win32 version with the system
1696    calls LoadLibrary/GetProcAddress.
1697    
1698    ----------------------------------------------------------------------
1699    Name: Matthias Blume (blume (at) tti - c (dot) org)
1700    Date: 2004/03/04 16:35:00 CST
1701    Tag: blume-20040304-intinf-fmt
1702    Description:
1703    
1704    Fixed problem with IntInf.fmt (sign would show up on the right instead
1705    of on the left for BIN, OCT, and HEX).
1706    
1707    ----------------------------------------------------------------------
1708    Name: Matthias Blume (blume (at) tti - c (dot) org)
1709    Date: 2004/03/04 11:25:00 CST
1710    Tag: blume-20040304-symlinks
1711    Description:
1712    
1713    Fixed problem with installer script (unix only) where bin/ml-yacc and
1714    friends pointed (via symlinks) to absolute locations instead of just
1715    .run-sml.  This was reported by Vesa A Norrman.
1716    
1717    ----------------------------------------------------------------------
1718    Name: Matthias Blume (blume (at) tti - c (dot) org)
1719    Date: 2004/02/13 14:50:00 CST
1720    Tag: Release_110_45
1721    Description:
1722    
1723    New working version (110.45). New bootfiles.
1724    
1725    ----------------------------------------------------------------------
1726    Name: Matthias Blume (blume (at) tti - c (dot) org)
1727    Date: 2004/01/26 15:15:15 CST
1728    Tag: blume-20040126-toplevel
1729    Description:
1730    
1731    Improved handling of exceptions at the interactive toplevel.
1732    
1733    ----------------------------------------------------------------------
1734    Name: Matthias Blume (blume (at) tti - c (dot) org)
1735    Date: 2004/01/26 11:25:00
1736    Tag: blume-20040126-app
1737    Description:
1738    
1739    Type of top-level "app" corrected.
1740    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1741    
1742    ----------------------------------------------------------------------
1743    Name: Matthias Blume (blume (at) tti - c (dot) org)
1744    Date: 2003/11/18 17:10 CST
1745    Tag: blume-20031118-basis-fiddle
1746    Description:
1747    
1748    - changed Timer interface to what might become the spec
1749    - POSIX_FLAGS -> BIT_FLAGS according to spec
1750    - some other minor discrepancies wrt. spec eliminated
1751    
1752    ----------------------------------------------------------------------
1753    Name: Matthias Blume (blume (at) tti - c (dot) org)
1754    Date: 2003/11/06 12:00:00 CST
1755    Tag: Release_110_44
1756    Description:
1757    
1758    New working version (110.44). New bootfiles.
1759    
1760    ----------------------------------------------------------------------
1761    Name: Matthias Blume (blume (at) tti - c (dot) org)
1762    Date: 2003/11/04 11:50:00 CST
1763    Tag: blume-20031104-move-libraries
1764    Description:
1765    
1766    Eliminated the "dont_move_libraries" directive in config/targets.
1767    (The mechanism was broken and could not be fixed easily.  Moreover,
1768    there does not seem to be any reason not to move all libraries into
1769    lib during installation.  I originally implemented this directive as a
1770    backward-compatibility feature when I first introduced the new CM.
1771    Now that things have been stable for a long time and going back to the
1772    old CM is not an option, there is no reason to keep it around.)
1773    
1774    ----------------------------------------------------------------------
1775    Name: Matthias Blume (blume (at) tti - c (dot) org)
1776    Date: 2003/11/03 16:00:00 CST
1777    Tag: blume-20031103-installdir
1778    Description:
1779    
1780    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1781    Richards for pointing out the problem and providing the solution.)
1782    
1783    ----------------------------------------------------------------------
1784    Name: Matthias Blume (blume (at) tti - c (dot) org)
1785    Date: 2003/10/01 17:05:00 CDT
1786    Tag: blume-20031001-lal-mlrisc
1787    Description:
1788    
1789    MLRISC bug fix from Lal.
1790    
1791    ----------------------------------------------------------------------
1792    Name: Matthias Blume (blume (at) tti - c (dot) org)
1793    Date: 2003/09/30 16:10:00 CDT
1794    Tag: blume-20030930-primio-bat
1795    Description:
1796    
1797    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1798    2. Improved .bat files (for Win32 port) to make things work under Win95.
1799       (thanks to Aaron S. Hawley for this one)
1800    
1801    ----------------------------------------------------------------------
1802    Name: Matthias Blume (blume (at) tti - c (dot) org)
1803    Date: 2003/09/26 16:05:00 CDT
1804    Tag: blume-20030926-wrappriv
1805    Description:
1806    
1807    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1808    
1809    ----------------------------------------------------------------------
1810    Name: Matthias Blume (blume (at) tti - c (dot) org)
1811    Date: 2003/09/26 15:00:00 CDT
1812    Tag: blume-20030926-110_43_3
1813    Description:
1814    
1815    - additional cleanup
1816    - version number bump, NEW BOOTFILES
1817    
1818    ----------------------------------------------------------------------
1819    Name: Matthias Blume (blume (at) tti - c (dot) org)
1820    Date: 2003/09/26 12:00:00 CDT
1821    Tag: blume-20030926-ppautoload
1822    Description:
1823    
1824    I modified the read-eval-print loop so that the autoloader gets
1825    invoked whenever the prettyprinter tries to look up a symbol that
1826    is not currently defined in the toplevel environment but which
1827    appears in CM's autoload registry.  As a result, we see far fewer of
1828    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1829    
1830    In addition to this I tried to clean up some pieces of the Basis
1831    implementation (e.g., Socket, Word8Array) in order to prevent other
1832    instances of these ?.Foo.Bar.xxx names from being printed.
1833    
1834    The mechanism that picks names for types still needs some work, though.
1835    (Right now it seems that if there is a type A.t which is defined to
1836    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1837    "?.B.u" although the perhaps more sensible solution would be to use
1838    "A.t" in this case.  In other words, the prettyprinter should follow
1839    a chain of DEFtycs not farther than there are corresponding toplevel
1840    names in the current environment.)
1841    
1842    ----------------------------------------------------------------------
1843    Name: Matthias Blume (blume (at) tti - c (dot) org)
1844    Date: 2003/09/24 16:31:00 CDT
1845    Tag: blume-20030924-installer
1846    Description:
1847    
1848    Another installer tweak:  All the ML code for the installer is now
1849    compiled during CMB.make and put into a little library called
1850    $smlnj/installer.cm.  The installation then simply invokes
1851    
1852       sml -m $smlnj/installer.cm
1853    
1854    and everything happens automagically.
1855    
1856    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1857    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1858            and BINDIR.
1859    
1860    The new scheme guarantees that the ML code responsible for the installation
1861    is in sync with the APIs of the main system.  Also, the installer is
1862    somewhat faster because the installer script is precompiled.
1863    
1864    ----------------------------------------------------------------------
1865    Name: Matthias Blume (blume (at) tti - c (dot) org)
1866    Date: 2003/09/24 15:35:00 CDT
1867    Tag: blume-20030924-synsock
1868    Description:
1869    
1870    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1871    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1872    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1873    implementations of non-blocking operations from CML's Socket
1874    structure.
1875    
1876    ----------------------------------------------------------------------
1877    Name: Matthias Blume (blume (at) tti - c (dot) org)
1878    Date: 2003/09/24 15:10:05 CDT
1879    Tag: blume-20030924-sockets
1880    Description:
1881    
1882    1. Fixed SOCKET API and implementation to match Basis spec.
1883       This required changing the internal representation of sockets to one
1884       that remembers (for each socket file descriptor) whether it is currently
1885       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1886       call is made only if the state actually needs to change).
1887    
1888    2. OS-specific details of sockets were moved into separate files, thus
1889       making it possible to unify the bulk of the socket implementations
1890       between Unix and Win32.
1891    
1892    3. CML's socket API changed accordingly.
1893       (Note that we need to remove non-blocking functions from this API
1894       since they are redundant in the case of CML!)
1895    
1896    4. CML's socket implementation now makes use of non-blocking functions
1897       provided by Basis, thus removing all OS-dependent code from this part
1898       of CML.
1899    
1900    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1901    
1902    ----------------------------------------------------------------------
1903    Name: Matthias Blume (blume (at) tti - c (dot) org)
1904    Date: 2003/09/22 12:10:00 CDT
1905    Tag: blume-20030922-110_43_2
1906    Description:
1907    
1908    Made a new interim version and bootfiles for developer's bootstrapping
1909    convenience.
1910    
1911       110.43.2 -- NEW BOOTFILES
1912    
1913    ----------------------------------------------------------------------
1914    Name: Matthias Blume (blume (at) tti - c (dot) org)
1915    Date: 2003/09/19 15:55:00 CDT
1916    Tag: blume-20030919-cmdir
1917    Description:
1918    
1919    1. new-install.sh -> install.sh
1920    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1921    3. tweaked installer so that another name instead of .cm can be chosen
1922       at install time (by setting the CM_DIR_ARC environment variable
1923       during installation); once installation is complete, the name is
1924       fixed
1925    
1926    ----------------------------------------------------------------------
1927    Name: Matthias Blume (blume (at) tti - c (dot) org)
1928    Date: 2003/09/18 16:00:00 CDT
1929    Tag: blume-20030918-110_43_1
1930    Description:
1931    
1932    Made a new interim version and bootfiles for developer's bootstrapping
1933    convenience.
1934    
1935       110.43.1 -- NEW BOOTFILES
1936    
1937    ----------------------------------------------------------------------
1938    Name: Matthias Blume (blume (at) tti - c (dot) org)
1939    Date: 2003/09/18 15:20:00 CDT
1940    Tag: blume-20030918-misc
1941    Description:
1942    
1943    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1944       this seems to be controversial at the moment) and used those in
1945       Posix.ProcEnv.times.
1946    
1947    2. Added Time.{from,to}Nanoseconds to Time.
1948    
1949    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1950       For example, fromLargeInt never needs to look at more than 3 "big
1951       digits" to get its 53 bits of precision.
1952    
1953    ----------------------------------------------------------------------
1954    Name: Matthias Blume (blume (at) tti - c (dot) org)
1955    Date: 2003/09/17 16:30:00 CDT
1956    Tag: blume-20030917-real32-slices
1957    Description:
1958    
1959    Added an entry to the primitive environment
1960    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1961    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1962    
1963    Removed some of the "magic" constants in real64.sml and replaced them
1964    with code that generates these values from their corresponding
1965    integer counterparts.
1966    
1967    Made all(?) the slice-related changes to the Basis and made everything
1968    compile again...
1969    
1970    ----------------------------------------------------------------------
1971    Name: Matthias Blume (blume (at) tti - c (dot) org)
1972    Date: 2003/09/15 17:45:00 CDT
1973    Tag: blume-20030915-rbase
1974    Description:
1975    
1976    Fixed bug in Real.fromLargeInt.
1977    
1978    ----------------------------------------------------------------------
1979    Name: Matthias Blume (blume (at) tti - c (dot) org)
1980    Date: 2003/09/13 18:11:00 CDT
1981    Tag: blume-20030913-libinstall
1982    Description:
1983    
1984    Minor bugfix in config/libinstall (set anchor with path to
1985    standalone tool after installing it, otherwise libraries that
1986    need ml-lex or ml-yacc won't compile the first time the installer
1987    runs).
1988    
1989    ----------------------------------------------------------------------
1990    Name: Matthias Blume (blume (at) tti - c (dot) org)
1991    Date: 2003/09/12 11:45:00 CDT
1992    Tag: blume-20030912-various
1993    Description:
1994    
1995     - fixed bug in Real.toLargeInt
1996     - fixed bug in Posix.ProcEnv.times
1997     - changed inputLine functions to return an option
1998     - minor installer improvements / bugfixes
1999     - changed default @SMLalloc parameter for x86/celeron to 64k
2000    
2001    ----------------------------------------------------------------------
2002    Name: Matthias Blume (blume (at) tti - c (dot) org)
2003    Date: 2003/09/09 22:00:00 CDT
2004    Tag: Release_110_43
2005    Description:
2006    
2007        New working release 110.43.  New bootfiles.
2008    
2009    ----------------------------------------------------------------------
2010    Name: Matthias Blume (blume (at) tti - c (dot) org)
2011    Date: 2003/09/09 19:20:00 CDT
2012    Tag: blume-20030909-installer
2013    Description:
2014    
2015    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
2016    Modified config/install.bat to take advantage of it.  Also modified
2017    config/install.sh (and called it config/new-install.sh) to take advantage
2018    of it on Unix systems.  (The SML code is (supposed to be) platform-
2019    independent.)
2020    
2021    The installer can now install everything under Win32
2022    as well as under *nix as long as it compiles.
2023    
2024    Other changes:
2025    
2026      - made CML compile again under Win32
2027      - made eXene compile under Win32 (by providing a fake structure UnixSock
2028        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
2029      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
2030        same as type int;  under Win32 it isn't
2031      - fixed some slice-related problems in the win32-specific parts of CML
2032      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
2033        (under Win32, the volume name is case-insensitive, and the
2034        OS.Path code compares volume names for equality)
2035    
2036    ----------------------------------------------------------------------
2037    Name: Matthias Blume (blume (at) tti - c (dot) org)
2038    Date: 2003/09/08 11:55:00 CDT
2039    Tag: blume-20030908-fullpath
2040    Description:
2041    
2042    Made Win32 version of OS.FileSys.fullPath return current directory
2043    when given an empty string.  This is what the spec says, and incidentally,
2044    CM depends on it.  (CM otherwise goes into an infinite loop in certain
2045    cases when presented with the name of a non-existing .cm file.)
2046    
2047    ----------------------------------------------------------------------
2048    Name: Matthias Blume (blume (at) tti - c (dot) org)
2049    Date: 2003/09/04 16:30:00 CDT
2050    Tag: blume-20030905-slices-etc
2051    Description:
2052    
2053      1. Changed interface to vectors and arrays in Basis to match
2054         (draft) Basis spec.
2055      2. Added signatures and implementations of slices according to
2056         Basis spec.
2057      3. Edited source code throughout the system to make it compile again
2058         under 1. and 2.  (In some cases code had to be added to have it
2059         match the new signatures.)
2060      4. MLRISC should be backward-compatible: the copies of the originals
2061         of files that needed to change under 3. were retained, the .cm files
2062         check the compiler version number and use old versions when
2063         appropriate.
2064      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
2065         dirstream -> string option (in accordance with Basis spec).
2066      6. When generating code that counts lines, ml-lex used function
2067         CharVector.foldli, taking advantage of its old interface.
2068         This has been replaced with the corresponding code from
2069         CharVectorSlice.  (html-lex must be re-lexed!)
2070      7. BitArray in smlnj-lib/Util has been extended/modified to match the
2071         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
2072      8. Removed temporary additions (fromInternal, toInternal) from the
2073         (now obsolete) IntInf in smlnj-lib/Util.
2074      9. Cleaned up structure Byte.
2075     10. Added localOffset, scan, and fromString to Date (according to spec).
2076         Cleaned/corrected implementation of Date.
2077         (Still need to check for correctness; implement better canonicalizeDate.)
2078     11. Added "scan" to signature IEEE_REAL.
2079     12. Some improvements to IntInf [in particular: efficiency-hack for
2080         mod and rem when second operand is 2 (for parity checks).]
2081     13. Changed representation of type Time.time, using a single IntInf.int
2082         value counting microseconds.  This considerably simplified the
2083         implementation of structure Time.  We now support negative time
2084         values; scan and fromString handle signs.
2085     14. Functor PrimIO now takes two additional arguments (VectorSlice and
2086         ArraySlice).
2087    
2088    ----------------------------------------------------------------------
2089    Name: Matthias Blume (blume (at) tti - c (dot) org)
2090    Date: 2003/08/28 17:00:00 CDT
2091    Tag: blume-20030828-intinf
2092    Description:
2093    
2094    This is a major update which comes with a version number bump
2095    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
2096    and an implementation of IntInf in the Basis.
2097    
2098    There are a fairly large number of related changes and updates throughout
2099    the system:
2100    
2101    Basis:
2102            - Implemented IntInf.
2103            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
2104            - Added some missing Real64 operations, most notably Real.toLargeInt.
2105            - Added FixedInt as a synonym for Int32.
2106    
2107    compiler:
2108            * Added support for a built-in intinf type.
2109              - literals
2110              - pattern matching
2111              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
2112              - overloading on literals and operations
2113    
2114            This required adding a primitive type intinf, some additional
2115            primops, and implementations for several non-trivial intinf
2116            operations in Core.  (The intinf type is completely abstract
2117            to the compiler; all operations get delegated back to the Core.)
2118    
2119            * Intinf equality is handled by polyequal.  However, the compiler
2120            does not print its usual warning in this case (since polyequal
2121            is the right thing to do there).
2122    
2123            * Improved the organization of structure InlineT.
2124    
2125            * A word about conversion primops:
2126            If conversions involving intinf do not cancel out during
2127            CPS contract, then the compiler must insert calls to Core functions.
2128            Since all core access must be resolved already during the FLINT
2129            translate phase, it would be too late a the time of CPS contract
2130            to add new Core calls.  For this reason, conversion primops
2131            for intinf carry two arguments:  1. the numeric argument that
2132            they are supposed to convert, and 2. the Core function that
2133            can help with this conversion if necessary.  If CPS contract
2134            eliminates a primop, then the associated Core function becomes
2135            dead and goes away.  Intinf conversion primops that do not get
2136            eliminated by CPS contract get rewritten into calls of their
2137            core functions by a separate, new phase.
2138    
2139    interactive system:
2140            - Control.Print.intinfDepth controls max length of intinf constants
2141            being printed. (Analogous to Control.Print.stringDepth.)
2142            - Cleanup in printutil and pputil: got rid of unused stuff and
2143            duplicates; replaced some of the code with code that makes better
2144            use of library functionality.
2145    
2146    CM:
2147            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
2148            are no longer silent.
2149    
2150    CKIT:
2151            Fixed mismatched uses of Int32 and LargeInt.  I always decided
2152            in favor of LargeInt -- which is now the same as IntInf.
2153            CKIT-knowledgable people should check whether this is what's
2154            intended and otherwise change things back to using Int32 or
2155            FixedInt.
2156    
2157    Throughout the code:
2158            Started using IntInf.int literals and built-in operations
2159            (e.g., comparison with 0) where this seems appropriate.
2160    
2161    
2162    ----------------------------------------------------------------------
2163    Name: Dave MacQueen (dbm@cs.uchicago.edu)
2164    Date: 2003/08/13 11:36:00 CDT
2165    Tag: dbm-20030813-mcz-merge1
2166    Description:
2167    
2168    Merging changes from the mcz-branch development branch into trunk.
2169    These changes involve replacement of the emulated old prettyprinter
2170    interface with direct use of the SML/NJ Lib PP library, and fixing
2171    of a couple of bugs (895, 1186) relating to error messages.  A new
2172    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
2173    has been added.
2174    
2175    ----------------------------------------------------------------------
2176    Name: Matthias Blume (blume (at) tti - c (dot) org)
2177    Date: 2003/08/11 15:45:00 CDT
2178    Tag: blume-20030811-windows
2179    Description:
2180    
2181    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
2182    
2183        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
2184    
2185    This patch restores SML/NJ's ability to run under win32.  There are a
2186    number of changes, including fixes for several bugs that had gone
2187    unnoticed until now:
2188    
2189      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
2190        trying to run the win32 version from within cygwin.)
2191    
2192      - There are a number of simple .bat scripts that substitute for their
2193        corresponding Unix shell-scripts.  (See below.)
2194    
2195      - The internals of ml-build have been modified slightly.  The main
2196        difference is that instead of calling ".link-sml" (or link-sml.bat)
2197        using OS.Process.system, the ML process delegates this task back
2198        to the script.  Otherwise problems arise in mixed environments such
2199        as Cygwin where scripts look and work like Unix scripts, but
2200        where OS.Process.system cannot run them.
2201    
2202      - In CM, the srcpath pickler used native pathname syntax -- which
2203        is incorrect in the case of cross-compilation.  The new pickle format
2204        is independent of platform-specific naming conventions.
2205    
2206      - Path configuration files (such as lib/pathconfig) can now choose
2207        between native and standard syntax.  Placing a line of the form
2208    
2209            standard!
2210    
2211        into the file causes all subsequent paths to be interpreted using
2212        CM standard pathname syntax (= Unix conventions); a line
2213    
2214            native!
2215    
2216        switches back to native style.  This was needed so that
2217        path config files can be written portably, see src/system/pathconfig.
2218    
2219      - Runtime system:
2220    
2221          - win32-filesys.c:  get_file_time and set_file_time now
2222            access modification time, not creation time.
2223    
2224          - I/O code made aware of new array representation.
2225    
2226          - Bug fixes in X86.prim.masm.
2227    
2228       - src/system/makeml made aware of win32. (For use under cygwin
2229            and other Unix-environments for windows.)
2230    
2231       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
2232         which caused BinIO.inputAll to fail consistently.
2233    
2234    .bat scripts:
2235    
2236       Windows .bat scripts assume that SMLNJ_HOME is defined.
2237    
2238       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
2239         applications (sml, ml-yacc, ml-lex).
2240       - ml-build.bat: analogous to ml-build.
2241       - config\install.bat:  Analogous to config/install.sh.  This requires
2242         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
2243         (nmake etc. must be on the path, and vcvars32 must have been run.)
2244         Moreover, sources for ml-lex and ml-yacc need to exist under src,
2245         and the bootfile hierarchy must have been unpacked under
2246         sml.boot.x86-win32.
2247         The script is very primitive and does a poor job at error checking.
2248         It only installs the base system, ml-lex, and ml-yacc.  No other
2249         libraries are being installed (i.e., you get only those that
2250         are part of the compiler.)
2251       - link-sml.bat:  analogous to .link-sml, but not currently used
2252    
2253    Unrelated bug fixes:
2254    
2255      - ml-nlffigen now exports structures ST_* corresponding to incomplete
2256        types.
2257      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
2258        otherwise.)
2259    
2260    ----------------------------------------------------------------------
2261    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
2262    Date: 2003/06/17
2263    Tag: macqueen-20030617-bug895
2264    Description:
2265    
2266    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
2267    Tag will be used for new development branch (mcz-branch) for
2268    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
2269    
2270    ----------------------------------------------------------------------
2271    Name: Matthias Blume (blume (at) tti - c (dot) org)
2272    Date: 2003/05/27 16:55:00 CDT
2273    Tag: blume-20030527-polyeq
2274    Description:
2275    
2276    Tried to eliminated most cases of polymorphic equality.
2277    
2278    ----------------------------------------------------------------------
2279    Name: Matthias Blume (blume (at) tti - c (dot) org)
2280    Date: 2003/05/21 17:45:00 CDT
2281    Tag: blume-20030517-complete
2282    Description:
2283    
2284    Two changes:
2285    
2286      1. Added a flag for controlling whether non-exhaustive bindings will
2287         be treated as errors (default is false).
2288      2. Cleaned up the *entire* source tree so that CMB.make goes through
2289         without a single non-exhaustive match- or bind warning.
2290    
2291    ----------------------------------------------------------------------
2292    Name: Matthias Blume (blume (at) tti - c (dot) org)
2293    Date: 2003/05/17 10:20:00 CDT
2294    Tag: blume-20030517-absyn
2295    Description:
2296    
2297    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
2298    
2299       This mainly affects the quality of error messages.  However, some
2300       of the code is now more straightforward than before. (Treatment of
2301       the above four constructs in translate.sml is much simpler than
2302       the "macro-expansion" that was going on before.  Plus, the mach-
2303       compiler no longer gets invoked just to be able to compile an
2304       if-expression.)
2305    
2306    2. The ErrorMsg.Error exception is now caught and absorbed by the
2307       interactive loop.
2308    
2309    ----------------------------------------------------------------------
2310    Name: Allen Leung
2311    Date: 2003/05/16 13:05:00 CDT
2312    Tag: leunga-20030516-cygwin-runtime
2313    Description:
2314    
2315        Ported the runtime system to cygwin, which uses the unix
2316    x86-unix bin files.  Missing/buggy features:
2317    
2318       o getnetbyname, getnetbyaddr: these functions seem to be missing in
2319         the Cygwin library.
2320       o Ctrl-C handling may be flaky.
2321       o Windows system calls and Windows I/O are not supported.
2322    
2323       A new set of binfiles is located at:
2324    
2325         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
2326    
2327       This is only needed for bootstrapping the cygwin version of smlnj.
2328       Other x86 versions can use the existing binfiles.
2329    
2330    ----------------------------------------------------------------------
2331    Name: Matthias Blume
2332    Date: 2003/04/08 15:42:00 CDT
2333    Tag: blume-20030408-listpair
2334    Description:
2335    
2336    1. Added a target 'mlrisc' to installer.
2337    
2338    2. Added missing elements to structure ListPair.
2339    
2340    ----------------------------------------------------------------------
2341    Name: Allen Leung
2342    Date: 2003/01/07 10:40:00 EST
2343    Tag: leunga-20030107-int-rem
2344    Description:
2345    
2346    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
2347    The arguments to the SUBL instruction were swapped.
2348    
2349    ----------------------------------------------------------------------
2350    Name: Matthias Blume
2351    Date: 2002/12/12 16:25:00 EST
2352    Tag: blume-20021212-risc-ra
2353    Description:
2354    
2355    Fixed a serious bug in the rewrite code for FP spilling/reloading that
2356    sent the RA into an infinite loop when floating point registers get
2357    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
2358    110.37 and 110.38.)
2359    There was another set of potential problems related to the handling of
2360    MLRISC annotations (but those did not yet cause real problems, apparently).
2361    
2362    ----------------------------------------------------------------------
2363    Name: Matthias Blume
2364    Date: 2002/12/06 22:40:00 EST
2365    Tag: blume-20021206-cm-fileid
2366    Description:
2367    
2368    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
2369    This fixes the problem of CM getting confused by files that suddenly
2370    change their identity (e.g., by getting unlinked and recreated by some
2371    text editor such as vi).  There might be a better/cheaper/cleaner way
2372    of doing this, but for now this will have to do.
2373    
2374    ----------------------------------------------------------------------
2375    Name: Matthias Blume
2376    Date: 2002/10/28 09:50:00 EST
2377    Tag: blume-20021028-typecheck
2378    Description:
2379    
2380    Exported structure Typecheck from $smlnj/viscomp/core.cm.
2381    
2382    ----------------------------------------------------------------------
2383    Name: Matthias Blume
2384    Date: 2002/10/17 09:10:00 EDT
2385    Tag: Release_110_42
2386    Description:
2387    
2388    In good old tradition, there has been a slight hiccup so that we have
2389    to patch 110.42 after the fact.   The old release tag has been replaced
2390    (see below).
2391    
2392    The change solves a problem with two competing approaches the
2393    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
2394    each other's way.
2395    
2396    This change only affects the runtime system code and the installer script.
2397    (No new bootfiles.)
2398    
2399    ----------------------------------------------------------------------
2400    Name: Matthias Blume
2401    Date: 2002/10/16 12:00:00 EDT
2402    Tag: Release_110_42_removed
2403    Description:
2404    
2405    New working release.  New bootfiles.
2406    
2407    ----------------------------------------------------------------------
2408    Name: Matthias Blume
2409    Date: 2002/10/10 13:10:00 EDT
2410    Tag: blume-20021010-ppc-divs
2411    Description:
2412    
2413    The mltree operator DIVS must be implemented with an overflow check on
2414    the PPC because the hardware indicates divide-by-zero using "overflow" as
2415    well.
2416    
2417    ----------------------------------------------------------------------
2418    Name: Matthias Blume
2419    Date: 2002/07/23 11:45:00
2420    Tag: blume-20020723-smlnj-home
2421    Description:
2422    
2423    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
2424    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
2425    CM_PATHCONFIG is also set), the path configuration file is assumed
2426    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
2427    the entire tree to some other place and everything will "just work".
2428    
2429    (Companion commands such as ml-build and ml-makedepend also sense this
2430    variable.)
2431    
2432    ----------------------------------------------------------------------
2433    Name: Matthias Blume
2434    Date: 2002/07/12 21:19:00 EDT
2435    Tag: blume-20020712-liveness
2436    Description:
2437    
2438    Exported two useful "step" functions from liveness module (MLRISC).
2439    
2440    ----------------------------------------------------------------------
2441    Name: Matthias Blume
2442    Date: 2002/07/05 16:00 EDT
2443    Tag: Release_110_41
2444    Description:
2445    
2446    New working release.  New bootfiles.
2447    
2448    ----------------------------------------------------------------------
2449    Name: Matthias Blume
2450    Date: 2002/07/05 10:25:00 EDT
2451    Tag: blume-20020705-btimp
2452    Description:
2453    
2454    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
2455    other clients can set up backtracing support.
2456    
2457    ----------------------------------------------------------------------
2458    Name: Matthias Blume
2459    Date: 2002/06/25 17:23:00 EDT
2460    Tag: blume-20020625-fpmax
2461    Description:
2462    
2463    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
2464    case.  (The sense of the isNaN test was reversed -- which made min and
2465    max always return their first argument.)
2466    
2467    ----------------------------------------------------------------------
2468    Name: Matthias Blume
2469    Date: 2002/06/11
2470    Tag: blume-20020611-unixpath
2471    Description:
2472    
2473    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
2474    
2475    ----------------------------------------------------------------------
2476    Name: Matthias Blume
2477    Date: 2002/06/10 16:35:00 EDT
2478    Tag: blume-20020610-ieeereal
2479    Description:
2480    
2481    I back-ported my implementation of IEEEReal.fromString from the
2482    idlbasis-devel branch so that we can test it.
2483    
2484    Another small change is that ppDec tries to give more information
2485    than just "<sig>" in the case of functors.  However, this code is
2486    broken in some mysterious way if the functor's body's signature
2487    has not been declared by ascription but gets inferred from the
2488    implementation.  This needs fixing...
2489    
2490    ----------------------------------------------------------------------
2491    Name: Matthias Blume
2492    Date: 2002/05/31
2493    Tag: blume-20020531-btrace-mode
2494    Description:
2495    
2496    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
2497    the wayside when I switched over to using Controls everywhere.)
2498    
2499    ----------------------------------------------------------------------
2500    Name: Lal George
2501    Date: 2002/05/23 12:21:40 EDT
2502    Tag: george-20020523-visual-labels
2503    Description:
2504    
2505    Labels are now displayed in the graphical output to make
2506    the fall-through and target blocks obvious.
2507    
2508    ----------------------------------------------------------------------
2509    Name: Matthias Blume
2510    Date: 2002/05/22 11:03:00 EDT
2511    Tag: blume-20020522-shrink
2512    Description:
2513    
2514    John tweaked yesterday's fix for 1131 to handle an out-of-memory
2515    situation that comes up when allocating huge arrays.
2516    
2517    ----------------------------------------------------------------------
2518    Name: Matthias Blume
2519    Date: 2002/05/21 16:00:00 EDT
2520    Tag: Release_110_40
2521    Description:
2522    
2523    New working release (110.40).  New bootfiles.
2524    
2525    [Also: John Reppy fixed GC bug 1131.]
2526    
2527    ----------------------------------------------------------------------
2528    Name: Matthias Blume
2529    Date: 2002/05/21 12:35:00 EDT
2530    Tag: blume-20020521-cmdoc
2531    Description:
2532    
2533    CM documentation update.
2534    
2535    ----------------------------------------------------------------------
2536    Name: Matthias Blume
2537    Date: 2002/05/21 10:55:00 EDT
2538    Tag: blume-20020521-misc
2539    Description:
2540    
2541    - John tweaked runtime to be silent on heap export (except when
2542      GC messages are on).
2543    - I added a few more things (cross-compiling versions of CMB) to
2544      config/preloads (as suggestions).
2545    
2546    ----------------------------------------------------------------------
2547    Name: Matthias Blume
2548    Date: 2002/05/20 22:25:00 EDT
2549    Tag: blume-20020520-controls
2550    Description:
2551    
2552    - Added ControlUtil structure to control-lib.cm.
2553    - Use it throughout.
2554    - Used Controls facility to define MLRISC controls (as opposed to
2555      registering MLRISC control ref cells with Controls after the fact)
2556    - Fixed messed-up controls priorities.
2557    
2558    * Removed again all the stuff from config/preloads that one wouldn't
2559      be able to preload at the time the initial heap image is built.
2560      (Many libraries, e.g., CML, do not exist yet at this time.  The
2561       only libraries that can be preloaded via config/preloads are those
2562       that come bundled with the bootfiles.)
2563    
2564    ----------------------------------------------------------------------
2565    Name: Matthias Blume
2566    Date: 2002/05/20 10:59:00 EDT
2567    Tag: blume-20020520-preloads
2568    Description:
2569    
2570    Added a lot of commented-out suggestions for things to be included
2571    in config/preloads.
2572    
2573    ----------------------------------------------------------------------
2574    Name: Allen Leung
2575    Date: 2002/05/18 14:20:00 EDT
2576    Tag: leunga-20020518-mdl
2577    Description:
2578    
2579    o Made the mdl tool stuff compile and run again.
2580    o I've disabled all the stuff that depends on RTL specifications; they
2581      are all badly broken anyway.
2582    
2583    ----------------------------------------------------------------------
2584    Name: Matthias Blume
2585    Date: 2002/05/17 16:49:00 EDT
2586    Tag: blume-20020517-controls
2587    Description:
2588    
2589    0. John Reppy made several modifications to the SML/NJ library.
2590       In particular, there is a shiny new controls-lib.cm.
2591    
2592    1. Pushed new controls interface through compiler so that everything
2593       compiles again.
2594    
2595    2. Added FormatComb and FORMAT_COMB to the CML version of the
2596       SML/NJ library (so that CML compiles again).
2597    
2598    3. Modified init scripts because XXX_DEFAULT environment variables
2599       are no longer with us.  (Boot-time initialization is now done
2600       using the same environment variables that are also used for
2601       startup-time initialization of controls.)
2602    
2603    ----------------------------------------------------------------------
2604    Name: Lal George
2605    Date: 2002/05/15 09:20:10 EDT
2606    Tag: george-20020515-pseudo-op-decls
2607    Description:
2608    
2609            All pseudo-ops emitted before the first segment declaration
2610    such as TEXT, DATA, and BSS directives are assumed to be global
2611    declarations and are emitted first in the assembly file. This is
2612    useful in a number of situations where one has pseudo-ops that are not
2613    specific to any segment, and also works around the constraint that one
2614    cannot have client pseudo-ops in the TEXT segment.
2615    
2616            Because no segment is associated with these declarations it is
2617    an error to allocate any space or objects before the first segment
2618    directive and an exception will be raised. However, we cannot make
2619    this check for client pseudo-ops.
2620    
2621            These top level declarations are a field in the CFG graph_info.
2622    In theory you can continue to add to this field after the CFG has been
2623    built -- provided you know what you are doing;-)
2624    
2625    ----------------------------------------------------------------------
2626    Name: Matthias Blume
2627    Date: 2002/05/13 16:40:00 EDT
2628    Tag: blume-20020513-pp-etc
2629    Description:
2630    
2631    A few minor bugfixes:
2632    
2633      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
2634        (Bogus printouts for redefined bindings still occur.  Compiler
2635        bug should no longer occur now.  We need to redo the prettyprinter
2636        from scratch.)
2637    
2638      - CM pathname printer now also adds escape sequences for ( and )
2639    
2640      - commend and docu fixes for ml-nlffi
2641    
2642    ----------------------------------------------------------------------
2643    Name: Matthias Blume
2644    Date: 2002/05/10 16:40:00 EDT
2645    Tag: blume-20020510-erg-textio
2646    Description:
2647    
2648    Applied the following bugfix provided by Emden Gansner:
2649    
2650        Output is corrupted when outputSubstr is used rather than output.
2651    
2652        The problem occurs when a substring
2653    
2654            ss = (s, dataStart, dataLen)
2655    
2656        where dataStart > 0, fills a stream buffer with avail bytes left.
2657        avail bytes of s, starting at index dataStart, are copied into the
2658        buffer, the buffer is flushed, and then the remaining dataLen-avail
2659        bytes of ss are copied into the beginning of the buffer. Instead of
2660        starting this copy at index dataStart+avail in s, the current code
2661        starts the copy at index avail.
2662    
2663        Fix:
2664        In text-io-fn.sml, change line 695 from
2665             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2666        to
2667             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2668    
2669    ----------------------------------------------------------------------
2670    Name: Matthias Blume
2671    Date: 2002/04/12 13:55:00 EDT
2672    Tag: blume-20020412-assyntax
2673    Description:
2674    
2675    1. Grabbed newer assyntax.h from the XFree86 project.
2676    2. Fiddled with how to compile X86.prim.asm without warnings.
2677    3. (Very) Minor cleanup in CM.
2678    
2679    ----------------------------------------------------------------------
2680    Name: Matthias Blume
2681    Date: 2002/04/01 (no joke!) 17:07:00 EST
2682    Tag: blume-20020401-x86div
2683    Description:
2684    
2685    Added full support for div/mod/rem/quot on the x86, using the machine
2686    instruction's two results (without clumsily recomputing the remainder)
2687    directly where appropriate.
2688    
2689    Some more extensive power-of-two support was added to the x86 instruction
2690    selector (avoiding expensive divs, mods, and muls where they can be
2691    replaced with cheaper shifts and masks).  However, this sort of thing
2692    ought to be done earlier, e.g., within the CPS optimizer so that
2693    all architectures benefit from it.
2694    
2695    The compiler compiles to a fixed point, but changes might be somewhat
2696    fragile nevertheless.  Please, report any strange things that you might
2697    see wrt. div/mod/quot/rem...
2698    
2699    ----------------------------------------------------------------------
2700    Name: Matthias Blume
2701    Date: 2002/03/29 17:22:00
2702    Tag: blume-20020329-div
2703    Description:
2704    
2705    Fixed my broken div/mod logic.  Unfortunately, this means that the
2706    inline code for div/mod now has one more comparison than before.
2707    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2708    The problem was with quotient = 0, because that alone does not tell
2709    us which way the rounding went.  One then has to look at whether
2710    remainder and divisor have the same sign...  :(
2711    
2712    Anyway, I replaced the bootfiles with fresh ones...
2713    
2714    ----------------------------------------------------------------------
2715    Name: Matthias Blume
2716    Date: 2002/03/29 14:10:00 EST
2717    Tag: blume-20020329-inlprims
2718    Description:
2719    
2720    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2721    
2722    Primops have changed. This means that the bin/boot-file formats have
2723    changed as well.
2724    
2725    To make sure that there is no confusion, I made a new version.
2726    
2727    
2728    CHANGES:
2729    
2730    * removed REMT from mltree (remainder should never overflow).
2731    
2732    * added primops to deal with divisions of all flavors to the frontend
2733    
2734    * handled these primops all the way through so they map to their respective
2735      MLRISC support
2736    
2737    * used these primops in the implementation of Int, Int32, Word, Word32
2738    
2739    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2740    
2741    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2742    
2743    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2744      floating point
2745    
2746    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2747      and Real (but Real.abs maps to a separate floating-point-only primop)
2748    
2749    
2750    TODO items:
2751    
2752    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2753      instructions because the machine instruction encoder cannot handle
2754      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2755      REM in the first place, and REM is supported by the x86 machine coder.)
2756    
2757    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2758      selection phase.  (The two can be streamlined because the hardware
2759      delivers both quotient and remainder at the same time anyway.)
2760    
2761    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2762      (Currently the behavior is inconsistent both across architectures and
2763      wrt. the draft Basis spec.)
2764    
2765    * Word8 should eventually be handled natively, too.
2766    
2767    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2768      as if there currently is no execution path that could trigger it in
2769      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2770      hold for things like multiplication and division.)
2771    
2772    ----------------------------------------------------------------------
2773    Name: Matthias Blume
2774    Date: 2002/03/27 16:27:00 EST
2775    Tag: blume-20020327-mlrisc-divisions
2776    Description:
2777    
2778    Added support for all four division operations (ML's div, mod, quot,
2779    and rem) to MLRISC.  In the course of doing so, I also rationalized
2780    the naming (no more annoying switch-around of DIV and QUOT), by
2781    parameterizing the operation by div_rounding_mode (which can be either
2782    DIV_TO_ZERO or DIV_TO_NEGINF).
2783    
2784    The generic MLTreeGen functor takes care of compiling all four
2785    operations down to only round-to-zero div.
2786    
2787    Missing pieces:
2788    
2789      * Doing something smarter than relying on MLTreeGen on architectures
2790        like, e.g., the x86 where hardware division delivers both quotient and
2791        remainder at the same time.  With this, the implementation of the
2792        round-to-neginf operations could be further streamlined.
2793    
2794      * Remove inlining support for div/mod/rem from the frontend and replace it
2795        with primops that get carried through to the backend.  Do this for all
2796        int and word types.
2797    
2798    ----------------------------------------------------------------------
2799    Name: Matthias Blume
2800    Date: 2002/03/25 17:25:00 EST
2801    Tag: blume-20020325-divmod
2802    Description:
2803    
2804    I improved (hopefully without breaking them) the implementation of Int.div,
2805    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2806    advantage of the following observations:
2807    
2808      Let  q = x quot y      r = x rem y
2809           d = x div  y      m = x mod y
2810    
2811    where "quot" is the round-to-zero version of integer division that
2812    hardware usually provides.  Then we have:
2813    
2814         r = x - q * y        where neither the * nor the - will overflow
2815         d = if q >= 0 orelse x = q * y then q else q - 1
2816                              where neither the * nor the - will overflow
2817         m = if q >= 0 orelse r = 0 then r else r + y
2818                              where the + will not overflow
2819    
2820    This results in substantial simplification of the generated code.
2821    The following table shows the number of CFG nodes and edges generated
2822    for
2823            fun f (x, y) = x OPER y
2824            (* with OPER \in div, mod, quot, rem *)
2825    
2826    
2827        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2828        --------------------------------------------------------
2829         div |         24 |         39 |         12 |         16
2830         mod |         41 |         71 |         12 |         16
2831        quot |          8 |         10 |          8 |         10
2832         rem |         10 |         14 |          8 |         10
2833    
2834    
2835    ----------------------------------------------------------------------
2836    Name: Matthias Blume
2837    Date: 2002/03/25 22:06:00 EST
2838    Tag: blume-20020325-cprotobug
2839    Description:
2840    
2841    Fixed a bug in cproto (c prototype decoder).
2842    
2843    ----------------------------------------------------------------------
2844    Name: Matthias Blume
2845    Date: 2002/03/25 16:00:00 EST
2846    Tag: blume-20020325-raw-primops
2847    Description:
2848    
2849    I did some cleanup to Allen's new primop code and
2850    replaced yesterday's bootfiles with new ones.
2851    (But they are stored in the same place.)
2852    
2853    ----------------------------------------------------------------------
2854    Name: Matthias Blume
2855    Date: 2002/03/24 22:40:00 EST
2856    Tag: blume-20020324-bootfiles
2857    Description:
2858    
2859    Made the bootfiles that Allen asked for.
2860    
2861    ----------------------------------------------------------------------
2862    Name: Allen Leung
2863    Date: 2002/03/23 15:50:00 EST
2864    Tag: leunga-20020323-flint-cps-rcc-primops
2865    Description:
2866    
2867      1. Changes to FLINT primops:
2868    
2869        (* make a call to a C-function;
2870         * The primop carries C function prototype information and specifies
2871         * which of its (ML-) arguments are floating point. C prototype
2872         * information is for use by the backend, ML information is for
2873         * use by the CPS converter. *)
2874      | RAW_CCALL of { c_proto: CTypes.c_proto,
2875                       ml_args: ccall_type list,
2876                       ml_res_opt: ccall_type option,
2877                       reentrant : bool
2878                     } option
2879       (* Allocate uninitialized storage on the heap.
2880        * The record is meant to hold short-lived C objects, i.e., they
2881        * are not ML pointers.  With the tag, the representation is
2882        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2883        *)
2884      | RAW_RECORD of {tag:bool,sz:int}
2885      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2886    
2887      2.  These CPS primops are now overloaded:
2888    
2889           rawload of {kind:numkind}
2890           rawstore of {kind:numkind}
2891    
2892          The one argument form is:
2893    
2894             rawload {kind} address
2895    
2896          The two argument form is:
2897    
2898             rawload {kind} [ml object, byte-offset]
2899    
2900      3. RAW_CCALL/RCC now takes two extra arguments:
2901    
2902         a. The first is whether the C call is reentrant, i.e., whether
2903            ML state should be saved and restored.
2904         b. The second argument is a string argument specifying the name of
2905            library and the C function.
2906    
2907         These things are currently not handled in the code generator, yet.
2908    
2909      4. In CProto,
2910    
2911         An encoding type of "bool" means "ml object" and is mapped into
2912         C prototype of PTR.  Note that "bool" is different than "string",
2913         even though "string" is also mapped into PTR, because "bool"
2914         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2915    
2916      5. Pickler/unpicker
2917    
2918         Changed to handle RAW_RECORD and newest RAW_CCALL
2919    
2920      6. MLRiscGen,
2921    
2922         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2923         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2924            in the file CodeGen/cpscompile/cps-c-calls.sml
2925    
2926      7. Added the conditional move operator
2927    
2928             condmove of branch
2929    
2930         to cps.  Generation of this is still buggy so it is currently
2931         disabled.
2932    
2933    ----------------------------------------------------------------------
2934    Name: Lal George
2935    Date: 2002/03/22 14:18:25 EST
2936    Tag: george-20020322-cps-branch-prob
2937    Description:
2938    
2939    Implemented the Ball-Larus branch prediction-heuristics, and
2940    incorporated graphical viewers for control flow graphs.
2941    
2942    Ball-Larus Heuristics:
2943    ---------------------
2944    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2945    
2946    By design it uses the Dempster-Shafer theory for combining
2947    probabilities.  For example, in the function:
2948    
2949        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2950    
2951    the ball-larus heuristics predicts that the n=0 is unlikely
2952    (OH-heuristic), and the 'then' branch is unlikely because of the
2953    RH-heuristic -- giving the 'then' branch an even lower combined
2954    probability using the Dempster-Shafer theory.
2955    
2956    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2957    probability of the 'then' branch because of the loop in the else
2958    branch.
2959    
2960    
2961    Graphical Viewing:
2962    ------------------
2963    I merely plugged in Allen's graphical viewers into the compiler. The
2964    additional code is not much. At the top level, saying:
2965    
2966            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2967    
2968    will display the graphical view of the control flow graph just before
2969    back-patching.  daVinci must be in your path for this to work. If
2970    daVinci is not available, then the default viewer can be changed
2971    using:
2972    
2973            Control.MLRISC.getString "viewer"
2974    
2975    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2976    course, these viewers must be in your path.
2977    
2978    The above will display the compilation unit at the level of clusters,
2979    many of which are small, boring, and un-interesting. Also setting:
2980    
2981            Control.MLRISC.getInt "cfg-graphical-view_size"
2982    
2983    will display clusters that are larger than the value set by the above.
2984    
2985    
2986    ----------------------------------------------------------------------
2987    Name: Matthias Blume
2988    Date: 2002/03/21 22:20:00 EST
2989    Tag: blume-20020321-kmp-bugfix
2990    Description:
2991    
2992    Changed the interface to the KMP routine in PreString and fixed
2993    a minor bug in one place where it was used.
2994    
2995    ----------------------------------------------------------------------
2996    Name: Allen Leung
2997    Date: 2002/03/21 20:30:00 EST
2998    Tag: leunga-20020321-cfg
2999    Description:
3000    
3001      Fixed a potential problem in cfg edge splitting.
3002    
3003    ----------------------------------------------------------------------
3004    Name: Allen Leung
3005    Date: 2002/03/21 17:15:00 EST
3006    Tag: leunga-20020321-x86-fp-cfg
3007    Description:
3008    
3009      1. Recoded the buggy parts of x86-fp.
3010    
3011         a. All the block reordering code has been removed.
3012            We now depend on the block placement phases to do this work.
3013    
3014         b. Critical edge splitting code has been simplified and moved into the
3015            CFG modules, as where they belong.
3016    
3017         Both of these were quite buggy and complex.  The code is now much, much
3018         simpler.
3019    
3020      2. X86 backend.
3021    
3022         a. Added instructions for 64-bit support.  Instruction selection for
3023            64-bit has not been committed, however, since that
3024            requires changes to MLTREE which haven't been approved by
3025            Lal and John.
3026    
3027         b. Added support for FUCOMI and FUCOMIP when generating code for
3028            PentiumPro and above.  We only generate these instructions in
3029            the fast-fp mode.
3030    
3031         c. Added cases for JP and JNP in X86FreqProps.
3032    
3033      3. CFG
3034    
3035         CFG now has a bunch of methods for edge splitting and merging.
3036    
3037      4. Machine description.
3038    
3039         John's simplification of MLTREE_BASIS.fcond broke a few machine
3040         description things:
3041    
3042         rtl-build.{sig,sml} and hppa.mdl fixed.
3043    
3044         NOTE: the machine description stuff in the repository is still broken.
3045               Again, I can't put my fixes in because that involves
3046               changes to MLTREE.
3047    
3048    ----------------------------------------------------------------------
3049    Name: Matthias Blume
3050    Date: 2002/03/20 15:55:00 EST
3051    Tag: blume-20020320-kmp
3052    Description:
3053    
3054    Implemented Knuth-Morris-Pratt string matching in PreString and used
3055    it for String.isSubstring, Substring.isSubstring, and
3056    Substring.position.
3057    
3058    (Might need some stress-testing.  Simple examples worked fine.)
3059    
3060    ----------------------------------------------------------------------
3061    Name: Matthias Blume
3062    Date: 2002/03/19 16:37:00 EST
3063    Tag: blume-20020319-witnesses
3064    Description:
3065    
3066    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
3067    
3068    This implements a generic mechanism for changing constness qualifiers
3069    anywhere within big C types without resorting to outright "casts".
3070    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
3071    modify the constness at the outermost level.)
3072    The implementation of "convert" is based on the idea of "witness"
3073    values -- values that are not used by the operation but whose types
3074    "testify" to their applicability.  On the implementation side, "convert"
3075    is simply a projection (returning its second curried argument).  With
3076    cross-module inlining, it should not result in any machine code being
3077    generated.
3078    
3079    ----------------------------------------------------------------------
3080    Name: Matthias Blume
3081    Date: 2002/03/15 16:40:00 EST
3082    Tag: blume-20020315-basis
3083    Description:
3084    
3085    Provided (preliminary?) implementations for
3086    
3087      {String,Substring}.{concatWith,isSuffix,isSubstring}
3088    
3089    and
3090    
3091      Substring.full
3092    
3093    Those are in the Basis spec but they were missing in SML/NJ.
3094    
3095    ----------------------------------------------------------------------
3096    Name: Matthias Blume
3097    Date: 2002/03/14 21:30:00 EST
3098    Tag: blume-20020314-controls
3099    Description:
3100    
3101    Controls:
3102    ---------
3103    
3104    1. Factored out the recently-added Controls : CONTROLS stuff and put
3105       it into its own library $/controls-lib.cm.  The source tree for
3106       this is under src/smlnj-lib/Controls.
3107    
3108    2. Changed the names of types and functions in this interface, so they
3109       make a bit more "sense":
3110    
3111          module -> registry
3112          'a registry -> 'a group
3113    
3114    3. The interface now deals in ref cells only.  The getter/setter interface
3115       is (mostly) gone.
3116    
3117    4. Added a function that lets one register an already-existing ref cell.
3118    
3119    5. Made the corresponding modifications to the rest of the code so that
3120       everything compiles again.
3121    
3122    6. Changed the implementation of Controls.MLRISC back to something closer
3123       to the original.  In particular, this module (and therefore MLRISC)
3124       does not depend on Controls.  There now is some link-time code in
3125       int-sys.sml that registers the MLRISC controls with the Controls
3126       module.
3127    
3128    CM:
3129    ---
3130    
3131      * One can now specify the lambda-split aggressiveness in init.cmi.
3132    
3133    ----------------------------------------------------------------------
3134    Name: Allen Leung
3135    Date: 2002/03/13 17:30:00 EST
3136    Tag: leunga-20020313-x86-fp-unary
3137    Description:
3138    
3139    Bug fix for:
3140    
3141    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
3142    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
3143    > - fun f(x,(y,z)) = Real.~ y;
3144    > [autoloading]
3145    > [autoloading done]
3146    >       fchsl   (%eax), 184(%esp)
3147    > Error: MLRisc bug: X86MCEmitter.emitInstr
3148    >
3149    > uncaught exception Error
3150    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
3151    
3152    The problem was that the code generator did not generate any fp registers
3153    in this case, and the ra didn't know that it needed to run the X86FP phase to
3154    translate the pseudo fp instruction.   This only happened with unary fp
3155    operators in certain situations.
3156    
3157    ----------------------------------------------------------------------
3158    Name: Matthias Blume
3159    Date: 2002/03/13 14:00:00 EST
3160    Tag: blume-20020313-overload-etc
3161    Description:
3162    
3163    1. Added _overload as a synonym for overload for backward compatibility.
3164       (Control.overloadKW must be true for either version to be accepted.)
3165    
3166    2. Fixed bug in install script that caused more things to be installed
3167       than what was requested in config/targets.
3168    
3169    3. Made CM aware of the (_)overload construct so that autoloading
3170       works.
3171    
3172    ----------------------------------------------------------------------
3173    Name: Matthias Blume
3174    Date: 2002/03/12 22:03:00 EST
3175    Tag: blume-20020312-url
3176    Description:
3177    
3178    Forgot to update BOOT and srcarchiveurl.
3179    
3180    ----------------------------------------------------------------------
3181    Name: Matthias Blume
3182    Date: 2002/03/12 17:30:00 EST
3183    Tag: blume-20020312-version110392
3184    Description:
3185    
3186    Yet another version number bump (because of small changes to the
3187    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
3188    
3189    Changes:
3190    
3191      The new pid generation scheme described a few weeks ago was overly
3192      complicated.  I implemented a new mechanism that is simpler and
3193      provides a bit more "stability":  Once CM has seen a compilation
3194      unit, it keeps its identity constant (as long as you do not delete
3195      those crucial CM/GUID/* files).  This means that when you change
3196      an interface, compile, then go back to the old interface, and
3197      compile again, you arrive at the original pid.
3198    
3199      There now also is a mechanism that instructs CM to use the plain
3200      environment hash as a module's pid (effectively making its GUID
3201      the empty string).  For this, "noguid" must be specified as an
3202      option to the .sml file in question within its .cm file.
3203      This is most useful for code that is being generated by tools such
3204      as ml-nlffigen (because during development programmers tend to
3205      erase the tool's entire output directory tree including CM's cached
3206      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
3207      revert to the old, broken behavior of SML/NJ, but in specific cases
3208      where there is no danger of interface confusion, its use is ok
3209      (I think).
3210    
3211      ml-nlffigen by default generates "noguid" annotations.  They can be
3212      turned off by specifying -guid in its command line.
3213    
3214    ----------------------------------------------------------------------
3215    Name: Lal George
3216    Date: 2002/03/12 12 14:42:36 EST
3217    Tag: george-20020312-frequency-computation
3218    Description:
3219    
3220    Integrated jump chaining and static block frequency into the
3221    compiler. More details and numbers later.
3222    
3223    ----------------------------------------------------------------------
3224    Name: Lal George
3225    Date: 2002/03/11 11 22:38:53 EST
3226    Tag: george-20020311-jump-chain-elim
3227    Description:
3228    
3229    Tested the jump chain elimination on all architectures (except the
3230    hppa).  This is on by default right now and is profitable for the
3231    alpha and x86, however, it may not be profitable for the sparc and ppc
3232    when compiling the compiler.
3233    
3234    The gc test will typically jump to a label at the end of the cluster,
3235    where there is another jump to an external cluster containing the actual
3236    code to invoke gc. This is to allow factoring of common gc invocation
3237    sequences. That is to say, we generate:
3238    
3239            f:
3240               testgc
3241               ja   L1      % jump if above to L1
3242    
3243            L1:
3244               jmp L2
3245    
3246    
3247    After jump chain elimination the 'ja L1' instructions is converted to
3248    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
3249    up being implemented in their long form (if L2 is far away) using:
3250    
3251            jbe     L3      % jump if below or equal to L3
3252            jmp     L2
3253         L3:
3254            ...
3255    
3256    
3257    For large compilation units L2  may be far away.
3258    
3259    
3260    ----------------------------------------------------------------------
3261    Name: Matthias Blume
3262    Date: 2002/03/11 13:30:00 EST
3263    Tag: blume-20020311-mltreeeval
3264    Description:
3265    
3266    A functor parameter was missing.
3267    
3268    ----------------------------------------------------------------------
3269    Name: Allen Leung
3270    Date: 2002/03/11 10:30:00 EST
3271    Tag: leunga-20020311-runtime-string0
3272    Description:
3273    
3274       The representation of the empty string now points to a
3275    legal null terminated C string instead of unit.  It is now possible
3276    to convert an ML string into C string with InlineT.CharVector.getData.
3277    This compiles into one single machine instruction.
3278    
3279    ----------------------------------------------------------------------
3280    Name: Allen Leung
3281    Date: 2002/03/10 23:55:00 EST
3282    Tag: leunga-20020310-x86-call
3283    Description:
3284    
3285       Added machine generation for CALL instruction (relative displacement mode)
3286    
3287    ----------------------------------------------------------------------
3288    Name: Matthias Blume
3289    Date: 2002/03/08 16:05:00
3290    Tag: blume-20020308-entrypoints
3291    Description:
3292    
3293    Version number bumped to 110.39.1.  NEW BOOTFILES!
3294    
3295    Entrypoints: non-zero offset into a code object where execution should begin.
3296    
3297    - Added the notion of an entrypoint to CodeObj.
3298    - Added reading/writing of entrypoint info to Binfile.
3299    - Made runtime system bootloader aware of entrypoints.
3300    - Use the address of the label of the first function given to mlriscGen
3301      as the entrypoint.  This address is currently always 0, but it will
3302      not be 0 once we turn on block placement.
3303    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
3304      with entry points) from mlriscGen.
3305    
3306    ----------------------------------------------------------------------
3307    Name: Allen Leung
3308    Date: 2002/03/07 20:45:00 EST
3309    Tag: leunga-20020307-x86-cmov
3310    Description:
3311    
3312       Bug fixes for CMOVcc on x86.
3313    
3314       1. Added machine code generation for CMOVcc
3315       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
3316       3. CMOVcc cannot have an immediate operand as argument.
3317    
3318    ----------------------------------------------------------------------
3319    Name: Matthias Blume
3320    Date: 2002/03/07 16:15:00 EST
3321    Tag: blume-20020307-controls
3322    Description:
3323    
3324    This is a very large but mostly boring patch which makes (almost)
3325    every tuneable compiler knob (i.e., pretty much everything under
3326    Control.* plus a few other things) configurable via both the command
3327    line and environment variables in the style CM did its configuration
3328    until now.
3329    
3330    Try starting sml with '-h' (or, if you are brave, '-H')
3331    
3332    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
3333    implements the underlying generic mechanism.
3334    
3335    The interface to some of the existing such facilities has changed somewhat.
3336    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
3337    (The getFoo interface is still there for backward-compatibility, but its
3338    use is deprecated.)
3339    
3340    The ml-build script passes -Cxxx=yyy command-line arguments through so
3341    that one can now twiddle the compiler settings when using this "batch"
3342    compiler.
3343    
3344    TODO items:
3345    
3346    We should go through and throw out all controls that are no longer
3347    connected to anything.  Moreover, we should go through and provide
3348    meaningful (and correct!) documentation strings for those controls
3349    that still are connected.
3350    
3351    Currently, multiple calls to Controls.new are accepted (only the first
3352    has any effect).  Eventually we should make sure that every control
3353    is being made (via Controls.new) exactly once.  Future access can then
3354    be done using Controls.acc.
3355    
3356    Finally, it would probably be a good idea to use the getter-setter
3357    interface to controls rather than ref cells.  For the time being, both
3358    styles are provided by the Controls module, but getter-setter pairs are
3359    better if thread-safety is of any concern because they can be wrapped.
3360    
3361    *****************************************
3362    
3363    One bug fix: The function blockPlacement in three of the MLRISC
3364    backpatch files used to be hard-wired to one of two possibilities at
3365    link time (according to the value of the placementFlag).  But (I
3366    think) it should rather sense the flag every time.
3367    
3368    *****************************************
3369    
3370    Other assorted changes (by other people who did not supply a HISTORY entry):
3371    
3372    1. the cross-module inliner now works much better (Monnier)
3373    2. representation of weights, frequencies, and probabilities in MLRISC
3374       changed in preparation of using those for weighted block placement
3375       (Reppy, George)
3376    
3377    ----------------------------------------------------------------------
3378    Name: Lal George
3379    Date: 2002/03/07 14:44:24 EST 2002
3380    Tag: george-20020307-weighted-block-placement
3381    
3382    Tested the weighted block placement optimization on all architectures
3383    (except the hppa) using AMPL to generate the block and edge frequencies.
3384    Changes were required in the machine properties to correctly
3385    categorize trap instructions. There is an MLRISC flag
3386    "weighted-block-placement" that can be used to enable weighted block
3387    placement, but this will be ineffective without block/edge
3388    frequencies (coming soon).
3389    
3390    
3391    ----------------------------------------------------------------------
3392    Name: Lal George
3393    Date: 2002/03/05 17:24:48 EST
3394    Tag: george-20020305-linkage-cluster
3395    
3396    In order to support the block placement optimization, a new cluster
3397    is generated as the very first cluster (called the linkage cluster).
3398    It contains a single jump to the 'real' entry point for the compilation
3399    unit. Block placement has no effect on the linkage cluster itself, but
3400    all the other clusters  have full freedom in the manner in which they
3401    reorder blocks or functions.
3402    
3403    On the x86 the typical linkage code that is generated is:
3404       ----------------------
3405            .align 2
3406       L0:
3407            addl    $L1-L0, 72(%esp)
3408            jmp     L1
3409    
3410    
3411            .align  2
3412       L1:
3413       ----------------------
3414    
3415    72(%esp) is the memory location for the stdlink register. This
3416    must contain the address of the CPS function being called. In the
3417    above example, it contains the address of  L0; before
3418    calling L1 (the real entry point for the compilation unit), it
3419    must contain the address for L1, and hence
3420    
3421            addl $L1-L0, 72(%esp)
3422    
3423    I have tested this on all architectures except the hppa.The increase
3424    in code size is of course negligible
3425    
3426    ----------------------------------------------------------------------
3427    Name: Allen Leung
3428    Date: 2002/03/03 13:20:00 EST
3429    Tag: leunga-20020303-mlrisc-tools
3430    
3431      Added #[ ... ] expressions to mlrisc tools
3432    
3433    ----------------------------------------------------------------------
3434    Name: Matthias Blume
3435    Date: 2002/02/27 12:29:00 EST
3436    Tag: blume-20020227-cdebug
3437    Description:
3438    
3439    - made types in structure C and C_Debug to be equal
3440    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
3441    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
3442    
3443    ----------------------------------------------------------------------
3444    Name: Matthias Blume
3445    Date: 2002/02/26 12:00:00 EST
3446    Tag: blume-20020226-ffi
3447    Description:
3448    
3449    1. Fixed a minor bug in CM's "noweb" tool:
3450       If numbering is turned off, then truly don't number (i.e., do not
3451       supply the -L option to noweb).  The previous behavior was to supply
3452       -L'' -- which caused noweb to use the "default" line numbering scheme.
3453       Thanks to Chris Richards for pointing this out (and supplying the fix).
3454    
3455    2. Once again, I reworked some aspects of the FFI:
3456    
3457       A. The incomplete/complete type business:
3458    
3459       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
3460         gone!
3461       - ML types representing an incomplete type are now *equal* to
3462         ML types representing their corresponding complete types (just like
3463         in C).  This is still safe because ml-nlffigen will not generate
3464         RTTI for incomplete types, nor will it generate functions that
3465         require access to such RTTI.   But when ML code generated from both
3466         incomplete and complete versions of the C type meet, the ML types
3467         are trivially interoperable.
3468    
3469         NOTE:  These changes restore the full generality of the translation
3470         (which was previously lost when I eliminated functorization)!
3471    
3472       B. Enum types:
3473    
3474       - Structure C now has a type constructor "enum" that is similar to
3475         how the "su" constructor works.  However, "enum" is not a phantom
3476         type because each "T enum" has values (and is isomorphic to
3477         MLRep.Signed.int).
3478       - There are generic access operations for enum objects (using
3479         MLRep.Signed.int).
3480       - ml-nlffigen will generate a structure E_foo for each "enum foo".
3481         * The structure contains the definition of type "mlrep" (the ML-side
3482         representation type of the enum).  Normally, mlrep is the same
3483         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
3484         then mlrep will be defined as a datatype -- thus facilitating
3485         pattern matching on mlrep values.
3486         ("-ec" will be suppressed if there are duplicate values in an
3487          enumeration.)
3488         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
3489         will be generated for each C enum constant xxx.
3490         * Conversion functions m2i and i2m convert between mlrep and
3491         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
3492         * Coversion functions c and ml convert between mlrep and "tag enum".
3493         * Access functions (get/set) fetch and store mlrep values.
3494       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
3495         enumerations are merged into one single enumeration represented by
3496         structure E_'.
3497    
3498    ----------------------------------------------------------------------
3499    Name: Allen Leung
3500    Date: 2002/02/25 04:45:00 EST
3501    Tag: leunga-20020225-cps-spill
3502    
3503    This is a new implementation of the CPS spill phase.
3504    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
3505    In case of problems, replace it with the old file spill.sml
3506    
3507    The current compiler runs into some serious performance problems when
3508    constructing a large record.  This can happen when we try to compile a
3509    structure with many items.  Even a very simple structure like the following
3510    makes the compiler slow down.
3511    
3512        structure Foo = struct
3513           val x_1 = 0w1 : Word32.int
3514           val x_2 = 0w2 : Word32.int
3515           val x_3 = 0w3 : Word32.int
3516           ...
3517           val x_N = 0wN : Word32.int
3518        end
3519    
3520    The following table shows the compile time, from N=1000 to N=4000,
3521    with the old compiler:
3522    
3523    N
3524    1000   CPS 100 spill                           0.04u  0.00s  0.00g
3525           MLRISC ra                               0.06u  0.00s  0.05g
3526              (spills = 0 reloads = 0)
3527           TOTAL                                   0.63u  0.07s  0.21g
3528    
3529    1100   CPS 100 spill                           8.25u  0.32s  0.64g
3530           MLRISC ra                               5.68u  0.59s  3.93g
3531              (spills = 0 reloads = 0)
3532           TOTAL                                   14.71u  0.99s  4.81g
3533    
3534    1500   CPS 100 spill                           58.55u  2.34s  1.74g
3535           MLRISC ra                               5.54u  0.65s  3.91g
3536              (spills = 543 reloads = 1082)
3537           TOTAL                                   65.40u  3.13s  6.00g
3538    
3539    2000   CPS 100 spill                           126.69u  4.84s  3.08g
3540           MLRISC ra                               0.80u  0.10s  0.55g
3541              (spills = 42 reloads = 84)
3542           TOTAL                                   129.42u  5.10s  4.13g
3543    
3544    3000   CPS 100 spill                           675.59u  19.03s  11.64g
3545           MLRISC ra                               2.69u  0.27s  1.38g
3546              (spills = 62 reloads = 124)
3547           TOTAL                                   682.48u  19.61s  13.99g
3548    
3549    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
3550           MLRISC ra                               4.96u  0.27s  2.72g
3551              (spills = 85 reloads = 170)
3552           TOTAL                                   2375.26u  57.21s  48.00g
3553    
3554    As you can see the old cps spill module suffers from some serious
3555    performance problem.  But since I cannot decipher the old code fully,
3556    instead of patching the problems up, I'm reimplementing it
3557    with a different algorithm.  The new code is more modular,
3558    smaller when compiled, and substantially faster
3559    (O(n log n) time and O(n) space).  Timing of the new spill module:
3560    
3561    4000  CPS 100 spill                           0.02u  0.00s  0.00g
3562          MLRISC ra                               0.25u  0.02s  0.15g
3563             (spills=1 reloads=3)
3564          TOTAL                                   7.74u  0.34s  1.62g
3565    
3566    Implementation details:
3567    
3568    As far as I can tell, the purpose of the CPS spill module is to make sure the
3569    number of live variables at any program point (the bandwidth)
3570    does not exceed a certain limit, which is determined by the
3571    size of the spill area.
3572    
3573    When the bandwidth is too large, we decrease the register pressure by
3574    packing live variables into spill records.  How we achieve this is
3575    completely different than what we did in the old code.
3576    
3577    First, there is something about the MLRiscGen code generator
3578    that we should be aware of:
3579    
3580    o MLRiscGen performs code motion!
3581    
3582       In particular, it will move floating point computations and
3583       address computations involving only the heap pointer to
3584       their use sites (if there is only a single use).
3585       What this means is that if we have a CPS record construction
3586       statement
3587    
3588           RECORD(k,vl,w,e)
3589    
3590       we should never count the new record address w as live if w
3591       has only one use (which is often the case).
3592    
3593       We should do something similar to floating point, but the transformation
3594       there is much more complex, so I won't deal with that.
3595    
3596    Secondly, there are now two new cps primops at our disposal:
3597    
3598     1. rawrecord of record_kind option
3599        This pure operator allocates some uninitialized storage from the heap.
3600        There are two forms:
3601    
3602         rawrecord NONE [INT n]  allocates a tagless record of length n
3603         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
3604                                     and initializes the tag.
3605    
3606     2. rawupdate of cty
3607          rawupdate cty (v,i,x)
3608          Assigns to x to the ith component of record v.
3609          The storelist is not updated.
3610    
3611    We use these new primops for both spilling and increment record construction.
3612    
3613     1. Spilling.
3614    
3615        This is implemented with a linear scan algorithm (but generalized
3616        to trees).  The algorithm will create a single spill record at the
3617        beginning of the cps function and use rawupdate to spill to it,
3618        and SELECT or SELp to reload from it.  So both spills and reloads
3619        are fine-grain operations.  In contrast, in the old algorithm
3620        "spills" have to be bundled together in records.
3621    
3622        Ideally, we should sink the spill record construction to where
3623        it is needed.  We can even split the spill record into multiple ones
3624        at the places where they are needed.  But CPS is not a good
3625        representation for global code motion, so I'll keep it simple and
3626        am not attempting this.
3627    
3628     2. Incremental record construction (aka record splitting).
3629    
3630        Long records with many component values which are simulatenously live
3631        (recall that single use record addresses are not considered to
3632         be live) are constructed with rawrecord and rawupdate.
3633        We allocate space on the heap with rawrecord first, then gradually
3634        fill it in with rawupdate.  This is the technique suggested to me
3635        by Matthias.
3636    
3637        Some restrictions on when this is applicable:
3638        1. It is not a VECTOR record.  The code generator currently does not handle
3639           this case. VECTOR record uses double indirection like arrays.
3640        2. All the record component values are defined in the same "basic block"
3641           as the record constructor.  This is to prevent speculative
3642           record construction.
3643    
3644    ----------------------------------------------------------------------
3645    Name: Allen Leung
3646    Date: 2002/02/22 01:02:00 EST
3647    Tag: leunga-20020222-mlrisc-tools
3648    
3649    Minor bug fixes in the parser and rewriter
3650    
3651    ----------------------------------------------------------------------
3652    Name: Allen Leung
3653    Date: 2002/02/21 20:20:00 EST
3654    Tag: leunga-20020221-peephole
3655    
3656    Regenerated the peephole files.  Some contained typos in the specification
3657    and some didn't compile because of pretty printing bugs in the old version
3658    of 'nowhere'.
3659    
3660    ----------------------------------------------------------------------
3661    Name: Allen Leung
3662    Date: 2002/02/19 20:20:00 EST
3663    Tag: leunga-20020219-mlrisc-tools
3664    Description:
3665    
3666       Minor bug fixes to the mlrisc-tools library:
3667    
3668       1.  Fixed up parsing colon suffixed keywords
3669       2.  Added the ability to shut the error messages up
3670       3.  Reimplemented the pretty printer and fixed up/improved
3671           the pretty printing of handle and -> types.
3672       4.  Fixed up generation of literal symbols in the nowhere tool.
3673       5.  Added some SML keywords to to sml.sty
3674    
3675    ----------------------------------------------------------------------
3676    Name: Matthias Blume
3677    Date: 2002/02/19 16:20:00 EST
3678    Tag: blume-20020219-cmffi
3679    Description:
3680    
3681    A wild mix of changes, some minor, some major:
3682    
3683    * All C FFI-related libraries are now anchored under $c:
3684        $/c.cm      --> $c/c.cm
3685        $/c-int.cm  --> $c/internals/c-int.cm
3686        $/memory.cm --> $c/memory/memory.cm
3687    
3688    * "make" tool (in CM) now treats its argument pathname slightly
3689      differently:
3690        1. If the native expansion is an absolute name, then before invoking
3691           the "make" command on it, CM will apply OS.Path.mkRelative
3692           (with relativeTo = OS.FileSys.getDir()) to it.
3693        2. The argument will be passed through to subsequent phases of CM
3694           processing without "going native".  In particular, if the argument
3695           was an anchored path, then "make" will not lose track of that anchor.
3696    
3697    * Compiler backends now "know" their respective C calling conventions
3698      instead of having to be told about it by ml-nlffigen.  This relieves
3699      ml-nlffigen from one of its burdens.
3700    
3701    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3702    
3703    * Export C_DEBUG and C_Debug from $c/c.cm.
3704    
3705    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3706      subtyping relationship between incomplete pointers and their complete
3707      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3708      with the convention of instantiating 'o with ('t, 'c) obj whenever
3709      the pointer target type is complete.  In the incomplete case, 'o
3710      will be instantiated with some "'c iobj" -- a type obtained by
3711      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3712    
3713      Operations that work on both incomplete and complete pointer types are
3714      typed as taking an 'o ptr while operations that require the target to
3715      be known are typed as taking some ('t, 'c) obj ptr.
3716    
3717      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3718      where void is an eqtype without any values.  This makes it possible
3719      to work on voidptr values using functions meant to operate on light
3720      incomplete pointers.
3721    
3722    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3723      been vastly simplified.
3724    
3725    ----------------------------------------------------------------------
3726    Name: Matthias Blume
3727    Date: 2002/02/19 10:48:00 EST
3728    Tag: blume-20020219-pqfix
3729    Description:
3730    
3731    Applied Chris Okasaki's bug fix for priority queues.
3732    
3733    ----------------------------------------------------------------------
3734    Name: Matthias Blume
3735    Date: 2002/02/15 17:05:00
3736    Tag: Release_110_39
3737    Description:
3738    
3739    Last-minute retagging is becoming a tradition... :-(
3740    
3741    This is the working release 110.39.
3742    
3743    ----------------------------------------------------------------------
3744    Name: Matthias Blume
3745    Date: 2002/02/15 16:00:00 EST
3746    Tag: Release_110_39-orig
3747    Description:
3748    
3749    Working release 110.39.  New bootfiles.
3750    
3751    (Update: There was a small bug in the installer so it wouldn't work
3752    with all shells.  So I retagged. -Matthias)
3753    
3754    ----------------------------------------------------------------------
3755    Name: Matthias Blume
3756    Date: 2002/02/15 14:17:00 EST
3757    Tag: blume-20020215-showbindings
3758    Description:
3759    
3760    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3761    the latter can be useful for exploring what bindings are available at
3762    the interactive prompt.  (The first function returns only the list
3763    of symbols that are really bound, the second prints those but also the
3764    ones that CM's autoloading mechanism knows about.)
3765    
3766    ----------------------------------------------------------------------
3767    Name: Matthias Blume
3768    Date: 2002/02/15 12:08:00 EST
3769    Tag: blume-20020215-iptrs
3770    Description:
3771    
3772    Two improvements to ml-nlffigen:
3773    
3774      1. Write files only if they do not exist or if their current contents
3775         do not coincide with what's being written.  (That is, avoid messing
3776         with the time stamps unless absolutely necessary.)
3777    
3778      2. Implement a "repository" mechanism for generated files related
3779         to "incomplete pointer types".   See the README file for details.
3780    
3781    ----------------------------------------------------------------------
3782    Name: Matthias Blume
3783    Date: 2002/02/14 11:50:00 EST
3784    Tag: blume-20020214-quote
3785    Description:
3786    
3787    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3788    because of the new and improved tag generation scheme.  (Thanks to Allen
3789    Leung for pointing it out.)
3790    
3791    ----------------------------------------------------------------------
3792    Name: Lal George
3793    Date: 2002/02/14 09:55:27 EST 2002
3794    Tag: george-20020214-isabelle-bug
3795    Description:
3796    
3797    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3798    of Isabelle on the x86.
3799    
3800    From Allen:
3801    -----------
3802     I've found the problem:
3803    
3804         in ra-core.sml, I use the counter "blocked" to keep track of the
3805         true number of elements in the freeze queue.  When the counter goes
3806         to zero, I skip examining the queue.  But I've messed up the
3807         bookkeeping in combine():
3808    
3809             else ();
3810             case !ucol of
3811               PSEUDO => (if !cntv > 0 then
3812                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3813                                        ^^^^^^^^^^^^^^^^^^^^^^^
3814                      moveu := mergeMoveList(!movev, !moveu)
3815                     )
3816                  else ();
3817    
3818         combine() is called to coalesce two nodes u and v.
3819         I think I was thinking that if the move counts of u and v are both
3820         greater than zero then after they are coalesced then one node is
3821         removed from the freeze queue.  Apparently I was thinking that
3822         both u and v are of low degree, but that's clearly not necessarily true.
3823    
3824    
3825    02/12/2002:
3826        Here's the patch.  HOL now compiles.
3827    
3828        I don't know how this impact on performance (compile
3829        time or runtime).  This bug caused the RA (especially on the x86)
3830        to go thru the potential spill phase when there are still nodes on the
3831        freeze queue.
3832    
3833    
3834    
3835    
3836    ----------------------------------------------------------------------
3837    Name: Matthias Blume
3838    Date: 2002/02/13 22:40:00 EST
3839    Tag: blume-20020213-fptr-rtti
3840    Description:
3841    
3842    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3843    updates.
3844    
3845    ----------------------------------------------------------------------
3846    Name: Matthias Blume
3847    Date: 2002/02/13 16:41:00 EST
3848    Tag: blume-20020213-cmlpq
3849    Description:
3850    
3851    Added new priority queue export symbols (which have just been added to
3852    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3853    not compile and the installer would choke.)
3854    
3855    ----------------------------------------------------------------------
3856    Name: Matthias Blume
3857    Date: 2002/02/13 16:15:00 EST
3858    Tag: blume-20020213-various
3859    Description:
3860    
3861    1. More tweaks to ml-nlffigen:
3862    
3863       - better internal datastructures (resulting in slight speedup)
3864       - "-match" option requires exact match
3865       - "localized" gensym counters (untagged structs/unions nested within
3866         other structs/unions or within typedefs get a fresh counter; their
3867         tag will be prefixed by a concatenation of their parents' tags)
3868       - bug fixes (related to calculation of transitive closure of types
3869         to be included in the output)
3870    
3871    2. Minor Basis updates:
3872    
3873       - added implementations for List.collate and Option.app
3874    
3875    ----------------------------------------------------------------------
3876    Name: Matthias Blume
3877    Date: 2002/02/11 15:55:00 EST
3878    Tag: blume-20020211-gensym
3879    Description:
3880    
3881    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3882    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3883    names (ML structure names that correspond to unnamed C structs, unions,
3884    and enums), so that separate runs of ml-nlffigen do not clash.
3885    
3886    ----------------------------------------------------------------------
3887    Name: Matthias Blume
3888    Date: 2002/02/11 12:05:00 EST
3889    Tag: blume-20020211-gensml
3890    Description:
3891    
3892    A quick fix for a problem with GenSML (in the pgraph-util library):
3893    Make generation of toplevel "local" optional.  (Strictly speaking,
3894    signature definitions within "local" are not legal SML.)
3895    
3896    Other than that: updates to INSTALL and cm/TODO.
3897    
3898    ----------------------------------------------------------------------
3899    Name: Matthias Blume
3900    Date: 2002/02/08 15:00:00 EST
3901    Tag: blume-20020208-uniquepid
3902    Description:
3903    
3904    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3905    
3906    1. The installer (config/install.sh) has gotten smarter:
3907    
3908         - Configuration options are a bit easier to specify now
3909           (in config/targets).
3910         - Bug in recognizing .tar.bz2 files fixed.
3911         - Installer automatically resolves dependencies between
3912           configuration options (e.g., if you ask for eXene, you will
3913           also get cml -- regardless whether you asked for it or not).
3914         - Installer can run in "quieter mode" by setting the environment
3915           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3916           "completely silent", though.
3917         - Build HashCons library as part of smlnj-lib.
3918    
3919    2. A new scheme for assigning persistent identifiers to compilation
3920       units (and, by extension, to types etc.) has been put into place.
3921       This fixes a long-standing bug where types and even dynamic values
3922       can get internally confused, thereby compromising type safety
3923       (abstraction) and dynamic correctness.  See
3924    
3925         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3926    
3927       for an example of how things could go wrong until now.
3928    
3929       The downside of the new scheme is that pids are not quite as
3930       persistent as they used to be: CM will generate a fresh pid
3931       for every compilation unit that it thinks it sees for the first
3932       time.  That means that if you compile starting from a clean, fresh
3933       source tree at two different times, you end up with different
3934       binaries.
3935    
3936       Cutoff recompilation, however, has not been compromised because
3937       CM keeps pid information in special caches between runs.
3938    
3939    ----------------------------------------------------------------------
3940    Name: Lal George
3941    Date: 2002/02/07 15:34:13 EST 2002
3942    Tag: <none>
3943    Description:
3944    
3945    Compilers that generate assembly code may produce  global labels
3946    whose value is resolved at link time. The various peephole optimization
3947    modules did not take this in account.
3948    
3949    TODO. The Labels.addrOf function should really return an option
3950    type so that clients are forced to deal with this issue, rather
3951    than an exception being raised.
3952    
3953    ----------------------------------------------------------------------
3954    Name: Lal George
3955    Date: 2002/02/06 13:55:02 EST
3956    Tag: george-20020206-ra-breakup
3957    Description:
3958    
3959    1. A bug fix from Allen.
3960    
3961        A typo causes extra fstp %st(0)'s to be generated at compensation
3962        edges, which might cause stack underflow traps at runtime.  This
3963        occurs in fft where there are extraneous fstps right before the 'into'
3964        trap instruction (in this case they are harmless since none of the
3965        integers overflow.)
3966    
3967    2. Pulled out various utility modules that were embedded in the modules
3968       of the register allocator. I need these modules for other purposes, but
3969       they are not complete enough to put into a library (just yet).
3970    ----------------------------------------------------------------------
3971    Name: Matthias Blume
3972    Date: 2002/01/31 16:05:00 EST
3973    Tag: blume-20020131-sparc-ccalls
3974    Description:
3975    
3976    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3977       of extra stack space by mistake.  Fixed.
3978    
3979    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3980    
3981    ----------------------------------------------------------------------
3982    Name: Allen Leung
3983    Date: 2002/01/30
3984    Tag: leunga-20020130-nowhere-bug-fix
3985    Description:
3986    
3987       MLRISC bug fixes:
3988       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3989       2. MachineInt.fromString was negating its value.
3990    
3991    ----------------------------------------------------------------------
3992    Name: Matthias Blume
3993    Date: 2002/01/29
3994    Tag: blume-20020129-INSTALL
3995    Description:
3996    
3997    - Added somewhat detailed installation instructions (file INSTALL).
3998    - Fixed curl-detection bug in config/install.sh.
3999    - It is now possible to select the URL getter using the URLGETTER
4000      environment variable:
4001    
4002          not set / "unknown"      --> automatic detection (script tries wget,
4003                                       curl, and lynx)
4004          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
4005                                       how to properly invoke them)
4006          other                    --> use $URLGETTER directly, it must take
4007                                       precisely two command-line arguments
4008                                       (source URL and destination file name)
4009    
4010    ----------------------------------------------------------------------
4011    Name: Matthias Blume
4012    Date: 2002/01/28
4013    Tag: blume-20020128-sparc-ccalls
4014    Description:
4015    
4016    - Fixed problem with calculation of "used" registers in sparc-c-calls.
4017    - Make use of the allocParam argument in sparc-c-calls.
4018    
4019    ----------------------------------------------------------------------
4020    Name: Matthias Blume
4021    Date: 2002/01/28
4022    Tag: blume-20020128-allocParam
4023    Description:
4024    
4025    John Reppy:  Changes c-calls API to accept client-callback for
4026    allocating extra stack space.
4027    me: Corresponding changes to mlriscGen (using a dummy argument that
4028        does not change the current behavior).
4029    
4030    ----------------------------------------------------------------------
4031    Name: Matthias Blume
4032    Date: 2002/01/28 12:00:00
4033    Tag: Release_110_38
4034    Description:
4035    
4036    This time for real!!!
4037    
4038    ----------------------------------------------------------------------
4039    Name: Matthias Blume
4040    Date: 2002/01/28 10:56:00 EST
4041    Tag: blume-20020128-retraction
4042    Description:
4043    
4044    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
4045       with blume-Release_110_38-retracted.)
4046    
4047    1. Fixed a problem with incorrect rounding modes in real64.sml.
4048       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
4049    
4050    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
4051       and enums fixed.  The naming of corresponding ML identifiers should
4052       now be consistent again.
4053    
4054    ----------------------------------------------------------------------
4055    Name: Allen Leung
4056    Date: 2002/01/27
4057    Tag: leunga-20020127-nowhere
4058    Description:
4059    
4060       Added a target called nowhere in the configuration scripts.
4061       Enabling this will build the MLRISC 'nowhere' tool (for translating
4062       programs with where-clauses into legal SML code) during installation.
4063    
4064    ----------------------------------------------------------------------
4065    Name: Matthias Blume
4066    Date: 2002/01/25 21:27:00 EST
4067    Tag: blume-Release_110_38-retracted
4068    Description:
4069    
4070    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
4071    
4072    README will be added later.
4073    
4074    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
4075    (Release_110_38) removed.  Reason: Last-minute bug fixes.
4076    
4077    ----------------------------------------------------------------------
4078    Name: Matthias Blume
4079    Date: 2002/01/25
4080    Tag: blume-20020125-ffi
4081    Description: