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

SCM Repository

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

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

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

revision 890, Thu Jul 19 20:38:56 2001 UTC revision 1867, Wed Nov 16 20:25:51 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: Matthias Blume (http://www.tti-c.org/blume/)
17    Date: 2005/11/16 14:30:00 CST
18    Tag: Release_110_57
19    Description:
20    
21    New working version (110.57).  NEW BOOTFILES!
22    
23    ----------------------------------------------------------------------
24    Name: Matthias Blume (http://www.tti-c.org/blume/)
25    Date: 2005/11/16 10:25:00 CST
26    Tag: blume-22051116-btrace
27    Description:
28    
29    Fixed problem with bogus exception message when using back-trace
30    facility.
31    
32    ----------------------------------------------------------------------
33    Name: Matthias Blume (http://www.tti-c.org/blume/)
34    Date: 2005/11/15 22:35:00 CST
35    Tag: blume-22051115-array2copy
36    Description:
37    
38    Added simple implementation of Array2.copy.  (Warning: mostly untested.)
39    
40    ----------------------------------------------------------------------
41    Name: Matthias Blume (http://www.tti-c.org/blume/)
42    Date: 2005/11/15 21:35:00 CST
43    Tag: blume-22051115-versiontool
44    Description:
45    
46    Reversed change to src/system/smlnj/internals/versiontool.cm.  This
47    file gets loaded as a tool -- by the equivalent of CM.make during the
48    run of CMB.make.  Thus, CMB's path configuration is meaningless for
49    it.  Instead, it has the status of "user code", so it should use
50    $/basis.cm to refer to the Basis library.  (At least that's true for
51    the purpose of bootstrapping the previous change.  In the future it
52    might make sense to have versiontool.cm refer to
53    $smlnj/basis/basis.cm, i.e., the version of the Basis that the
54    compiler itself uses.)
55    
56    Also patched src/system/testml to have it activate those extra anchor
57    bindings in config/extrapathconfig.
58    
59    ----------------------------------------------------------------------
60    Name: Matthias Blume (http://www.tti-c.org/blume/)
61    Date: 2005/11/15 16:35:00 CST
62    Tag: blume-22051115-libraries
63    Description:
64    
65    This change affects the way the following libraries are tied into
66    the system:
67    
68       $/basis.cm
69       $/smlnj-lib.cm
70       $/pp-lib.cm
71       $/controls-lib.cm
72       $/html-lib.cm
73       $/ml-yacc-lib.cm
74    
75    These libraries are now internally (as seen from the source code
76    of the implementation itself) known by the following names:
77    
78       $smlnj/basis/basis.cm
79       $smlnj/smlnj-lib/smlnj-lib.cm
80       $smlnj/smlnj-lib/pp-lib.cm
81       $smlnj/smlnj-lib/controls-lib.cm
82       $smlnj/smlnj-lib/html-lib.cm
83       $smlnj/ml-yacc/ml-yacc-lib.cm
84    
85    This makes it possible to work with code that requires different
86    versions of these libraries, and which refers to these libraries using
87    their "default" names (i.e., the first set of names above).  In other
88    words, one can un-define or re-define those default names without
89    compromising the proper functioning of the compiler itself.
90    
91    A similar procedure had already been performed for several of the
92    MLRISC libraries that are linked into the compiler.  I did some
93    cleanup on this code.
94    
95    A new file in the config directory (named extrapathconfig) is
96    responsible for setting up path anchors that the compiler itself does
97    not need, but that are typically required by user code.
98    
99    ----------------------------------------------------------------------
100    Name: Matthias Blume
101    Date: 2005/11/07 17:05:00 CST
102    Tag: blume-20051107-slice
103    Description:
104    
105    Fixed erroneous out-of-bounds test in the "update" function of
106    various *ArraySlice modules.  (Thanks to Vesa A. Norrman for pointing
107    out the problem.)
108    
109    Pushed some Basis changes through ML-Lex, CML, and eXene.
110    
111    ----------------------------------------------------------------------
112    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
113    Date: 2005/11/07
114    Description:
115    Fixed a Basis incompatibility: The depreciated function Substring.all
116    was removed (use Substring.full instead).
117    
118    ----------------------------------------------------------------------
119    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
120    Date: 2005/11/05
121    Description:
122    Tweaked interval set API in SML/NJ library; see the CHANGES file for details.
123    
124    ----------------------------------------------------------------------
125    Name: Matthias Blume (http://www.tti-c.org/blume/)
126    Date: 2005/11/02 16:25:00 CST
127    Tag: blume-20051102-hex
128    Description:
129    
130    * Runtime system bootstrap code now accepts hex digits in BOOTLIST
131      in either upper case or lower case format.
132    * Pushed changes to names of Pack<N>{Big,Little} structures through
133      CML and eXene.
134    
135    ----------------------------------------------------------------------
136    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
137    Date: 2005/11/02
138    Description:
139    Fixed a Basis incompatibility: Pack<N>{Big,Little} structures should
140    be called PackWord<N>{Big,Little}.
141    
142    ----------------------------------------------------------------------
143    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
144    Date: 2005/10/28
145    Description:
146    Fixed a minor Basis incompatibility: hex digits should be upper case.
147    
148    ----------------------------------------------------------------------
149    Name: Matthias Blume (blume (at) tti - c (dot) org)
150    Date: 2005/10/25 16:00:00 CDT
151    Tag: Release_110_56
152    Description:
153    
154    New working version (110.56).  NEW BOOTFILES!
155    
156    ----------------------------------------------------------------------
157    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
158    Date: 2005/10/25
159    Description:
160    Added interval sets to utility library (signatures INTERVAL_DOMAIN
161    and INTERVAL_SET, and functor IntervalSetFn).
162    
163    ----------------------------------------------------------------------
164    Name: John Reppy (http://www.cs.uchicago.edu/~jhr)
165    Date: 2005/10/14
166    Description:
167    Add Zhong Shao's fix for datatype equality functions.
168    
169    ----------------------------------------------------------------------
170    Name: Allen Leung (leunga (at) reservoir (dot) com)
171    Date: 2005/10/14 12:30:00 EST
172    Tag: leunga-20051014-x86-spill
173    Description:
174    
175    Bug fix a bug found by Carl Hauser.
176    There was a typo in the reload code for FCMP in x86SpillInstr.sml.
177    
178    ----------------------------------------------------------------------
179    Name: Allen Leung (leunga (at) reservoir (dot) com)
180    Date: 2005/10/14 11:15:00 EST
181    Tag: leunga-20051014-x86-asm
182    Description:
183    
184    Removed some debugging code in file x86Asm.sml.
185    The function emit_operand was printing out debugging output.
186    
187    ----------------------------------------------------------------------
188    Name: John Reppy
189    Date: 2005/07/27
190    Tag:
191    Description:
192    
193    Fixed ml-lex to recognize "\r" as representing carriage return.
194    
195    ----------------------------------------------------------------------
196    Name: John Reppy
197    Date: 2005/07/27
198    Tag:
199    Description:
200    
201    Fixed ml-yacc to work on files with non-native end-of-line
202    encodings (e.g., Windows text file on a Unix system).
203    
204    ----------------------------------------------------------------------
205    Name: John Reppy
206    Date: 2005/07/20
207    Tag:
208    Description:
209    
210    Added changes from Dominic Evans (oldmanuk (at) gmail (dot) com)
211    to support HPUX 11.
212    
213    ----------------------------------------------------------------------
214    Name: John Reppy
215    Date: 2005/07/06
216    Tag:
217    Description:
218    
219    Changes to the SML/NJ library.  See smlnj-lib/CHANGES for details.
220    
221    ----------------------------------------------------------------------
222    Name: Matthias Blume (blume (at) tti - c (dot) org)
223    Date: 2005/07/06 00:45:00 CDT
224    Tag: blume-20050706-slice-copy
225    Description:
226    
227    Fixed reversed logic for deciding whether to "copy up" or "copy down"
228    in *-array-slice.sml.
229    
230    ----------------------------------------------------------------------
231    Name: Allen Leung (leunga (at) reservoir (dot) com)
232    Date: 2005/05/31 17:00:00 EST
233    Tag: leunga-20050531-cygwin-fault-2
234    Description:
235    
236    A typo in the cygwin code fixed.
237    
238    ----------------------------------------------------------------------
239    Name: Allen Leung (leunga (at) reservoir (dot) com)
240    Date: 2005/05/31 16:47:00 EST
241    Tag: leunga-20050531-cygwin-fault
242    Description:
243    
244    Updated Cygwin's fault/signal handling to match the Windows version.
245    Updated the export list.
246    
247    ----------------------------------------------------------------------
248    Name: Matthias Blume (blume (at) tti - c (dot) org)
249    Date: 2005/05/18 13:35:00 CDT
250    Tag: Release_110_54
251    Description:
252    
253    New working version (110.54).  NEW BOOTFILES!
254    
255    ----------------------------------------------------------------------
256    Name: Matthias Blume (blume (at) tti - c (dot) org)
257    Date: 2005/05/18 11:58:00 CDT
258    Tag: blume-20050518-installer
259    Description:
260    
261    Added support scripts for Mac OS X PackageMaker and modified
262    config/install.sh so that it supports re-dumping a heap image
263    after customization.
264    
265    ----------------------------------------------------------------------
266    Name: Matthias Blume (blume (at) tti - c (dot) org)
267    Date: 2005/05/18 10:55:00 CDT
268    Tag: blume-20050518-realdiv-noovld
269    Description:
270    
271    Un-overloaded / to work around bug in overloading resolution code.
272    
273    ----------------------------------------------------------------------
274    Name: Matthias Blume (blume (at) tti - c (dot) org)
275    Date: 2005/05/16 23:50:00 CDT
276    Tag: blume-20050516-redump-heap
277    Description:
278    
279    Added mechanism for re-creating a heap file for the interactive system
280    after configuration variables have been changed.
281    
282       CM.redump_heap : string -> unit
283    
284    This is much like SMLofNJ.exportML, but starting from the resulting
285    heap does not return to the caller of CM.redump_heap but
286    restarts the interactive system from scratch.  The original call of
287    CM.redump_heap does not return but ends the interactive session.
288    Thus, CM.redump_heap is a lot like SMLofNJ.exportFn.
289    
290    Internally, redump_heap winds the dynamic execution context back to
291    the point where the original heap image was created and re-executes
292    the heap image generation code in the boot code.
293    
294    ----------------------------------------------------------------------
295    Name: Matthias Blume (blume (at) tti - c (dot) org)
296    Date: 2005/05/09 21:55:00 CDT
297    Tag: blume-20050509-word64
298    Description:
299    
300    Added a hack to the existing hack known as Word64 to make fromString
301    behave correctly.  I am still not sure whether Word64.scan will work
302    as specified with respect to the interaction of radix and prefix.
303    
304    ----------------------------------------------------------------------
305    Name: Allen Leung (leunga (at) reservoir (dot) com)
306    Date: 2005/05/04 11:50:00 EST
307    Tag: leunga-20050504-checkgc
308    Description:
309    
310    Added a gc protocol checking phase.  This phase is enabled with
311    the flag "check-gc".  "debug-check-gc" turns on the verbose mode.
312    
313    ----------------------------------------------------------------------
314    Name: Matthias Blume (blume (at) tti - c (dot) org)
315    Date: 2005/05/04 10:45:00 CDT
316    Tag: blume-20050504-intinf
317    Description:
318    
319    Fixed a bug in the implementation of div and mod for IntInf.
320    Thanks to Neophytos Michael for reporting the problem.
321    
322    ----------------------------------------------------------------------
323    Name: Matthias Blume (blume (at) tti - c (dot) org)
324    Date: 2005/05/04 10:35:00 CDT
325    Tag: blume-20050504-join
326    Description:
327    
328    Added a "join" combinator to the ParserComb module in smlnj-lib.cm.
329    
330    ----------------------------------------------------------------------
331    Name: Matthias Blume (blume (at) tti - c (dot) org)
332    Date: 2005/02/28 23:40:00 CST
333    Tag: blume-20050228-mVar
334    Description:
335    
336    Fixed serious bug (brown paper bag variety) in new implementation of
337    structure Atom in CML.  (I had accidentally used a mailbox instead of
338    an mvar, leaving the door open for races.)
339    
340    ----------------------------------------------------------------------
341    Name: Matthias Blume (blume (at) tti - c (dot) org)
342    Date: 2005/02/25 15:00:00 CST
343    Tag: Release_110_53
344    Description:
345    
346    New working version (110.53).  NEW BOOTFILES!
347    
348    ----------------------------------------------------------------------
349    Name: Matthias Blume (blume (at) tti - c (dot) org)
350    Date: 2005/02/25 14:50:00 CST
351    Tag: blume-20050225-susp
352    Description:
353    
354    Brought back SMLofNJ.Susp.  The underlying suspension type is the one
355    implemented in Core -- which means that it is the same as the one used
356    by the lazy extension.
357    
358    ----------------------------------------------------------------------
359    Name: Matthias Blume (blume (at) tti - c (dot) org)
360    Date: 2005/02/24 16:50:00 CST
361    Tag: blume-20050224-cml-atom
362    Description:
363    
364    Simpler and at the same time more general implementation of structure
365    Atom in CML.
366    
367    ----------------------------------------------------------------------
368    Name: Matthias Blume (blume (at) tti - c (dot) org)
369    Date: 2005/02/15 17:35:00 CST
370    Tag: blume-20050215-tools
371    Description:
372    
373    Created new "tools" directory under "src" and moved "TraceDebugProf"
374    there.
375    
376    ----------------------------------------------------------------------
377    Name: Matthias Blume (blume (at) tti - c (dot) org)
378    Date: 2005/02/10 17:55:00 CST
379    Tag: blume-20050210-longlong
380    Description:
381    
382    Implemented "long long" arguments and results for NLFFI.  (Only the
383    PPC/MacOS implementation is complete, the other backends still need to
384    be updated.)
385    
386    ----------------------------------------------------------------------
387    Name: Matthias Blume (blume (at) tti - c (dot) org)
388    Date: 2005/01/24 17:40:00 CST
389    Tag: blume-20050124-mlyacc
390    Description:
391    
392    Minor cleanup in ML-Yacc rule printing mechanism.  This should fix a
393    problem with certain "as" patterns which previously got rendered
394    using incorrect syntax.
395    
396    ----------------------------------------------------------------------
397    Name: Matthias Blume (blume (at) tti - c (dot) org)
398    Date: 2005/01/18 12:00:00 CST
399    Tag: blume-20050118-profile
400    Description:
401    
402    Made time profiling code (interrupt handler) in runtime system aware
403    of new array representation.
404    
405    ----------------------------------------------------------------------
406    Name: Matthias Blume (blume (at) tti - c (dot) org)
407    Date: 2005/01/14 18:00:00 CST
408    Tag: blume-20050114-heap2exec
409    Description:
410    
411    Implemented new (but still experimental) heap2exec facility.  This is
412    tested under Mac OS X and should work under Linux (will test shortly).
413    It will probably also work on the Sparc (will test some time later).
414    
415      - removed old "HACKED_STANDALONE" hack from runtime
416    
417    To be able to test this, uncomment the request for "heap2asm" in
418    config/targets prior to installation.  (Notice that this is different
419    from "heap2exec" mentioned below.  Not a typo.)
420    
421    To perform an actual test, run the command
422    
423       $ bin/heap2exec heapfile execfile
424    
425    (You can put heap2exec on your shell's path.)
426    
427    For example, run
428    
429      $ bin/heap2exec bin/.heap/ml-yacc.ppc-darwin mly
430    
431    This will create a standalone executable called "mly" which you can
432    then invoke directly as a command.
433    
434    ----------------------------------------------------------------------
435    Name: Matthias Blume (blume (at) tti - c (dot) org)
436    Date: 2005/01/07 11:44:00 CST
437    Tag: blume-20050107-mlstring
438    Description:
439    
440    fixed off-by-one error in ML_STRING macro (globals.c)
441    
442    ----------------------------------------------------------------------
443    Name: Matthias Blume (blume (at) tti - c (dot) org)
444    Date: 2004/12/23 18:00:00 CST
445    Tag: blume-20041223-santa
446    Description:
447    
448    Made ml-build script "smarter" (but only very little).
449    
450    ----------------------------------------------------------------------
451    Name: Matthias Blume (blume (at) tti - c (dot) org)
452    Date: 2004/12/21 15:05:00 CST
453    Tag: blume-20041221-longlong
454    Description:
455    
456    * Implemented access to signed and unsigned long long data in NLFFI.
457    (The parameter-passing part of the picture has not complete. But
458    data structure access seems to work.)
459    
460    * Fixed CM's incorrect assumption that the PPC is little-endian.
461      (On the Mac, it is big-endian.  And that's currently our only
462      PPC platform.)
463    
464    ----------------------------------------------------------------------
465    Name: Matthias Blume (blume (at) tti - c (dot) org)
466    Date: 2004/12/21 12:50:00 CST
467    Tag: blume-20041221-memory
468    Description:
469    
470    Some cleanup in the $c/memory.cm library: separated some concerns by
471    moving allocation code and memory access code each into their own
472    files.
473    
474    ----------------------------------------------------------------------
475    Name: Allen Leung (leunga (at) reservoir (dot) com)
476    Date: 2004/12/17 16:12:00 EST
477    Tag: leunga-20041217-cygwin-smlnj-home
478    Description:
479    
480        The Unix I/O library of SML/NJ on cygwin does not understand
481    Windows style pathname, so problems arise when SMLNJ_HOME is set
482    to a Windows style pathname.  _run-sml now converts SMLNJ_HOME
483    to a POSIX pathname on cygwin.
484    
485    ----------------------------------------------------------------------
486    Name: Matthias Blume (blume (at) tti - c (dot) org)
487    Date: 2004/12/16 13:35:00 CST
488    Tag: Release_110_52
489    Description:
490    
491    Last-minute changes incorporated into 110.52.  Release tag moved.
492    
493    The changes:
494       - HashString.hashString' -> HashString.hashSubstring
495       - bug fix in UnivariateStats
496    
497    ----------------------------------------------------------------------
498    Name: Matthias Blume (blume (at) tti - c (dot) org)
499    Date: 2004/12/15 23:40:00 CST
500    Tag: blume-20041215-hashSubstring
501    Description:
502    
503    - HashString.hashString' -> HashString.hashSubstring
504    - corresponding changes in atom.sml
505    - "de-compressed" (aka. un-obfuscated) code for UnivariateStats and
506      added some comments
507    
508    ----------------------------------------------------------------------
509    Name: Matthias Blume (blume (at) tti - c (dot) org)
510    Date: 2004/12/15 15:30:00 CST
511    Tag: (Release_110_52)
512    Description:
513    
514    New working version (110.52).  NEW BOOTFILES!
515    
516    ----------------------------------------------------------------------
517    Name: Matthias Blume (blume (at) tti - c (dot) org)
518    Date: 2004/12/15 12:45:00 CST
519    Tag: blume-20041215-spaces
520    Description:
521    
522    More on the space problem (this time for Win32).
523    
524    ----------------------------------------------------------------------
525    Name: Matthias Blume (blume (at) tti - c (dot) org)
526    Date: 2004/12/14 17:30:00 CST
527    Tag: blume-20041214-spaces
528    Description:
529    
530    Hacked some of the scripts (in particular: the installer) to cope with
531    spaces in filenames a bit better.  But beware: the current "solution"
532    is likely still full of bugs and inherently incomplete.  (We need to
533    do away with those shell scripts for a comprehensive solution.)
534    
535    ----------------------------------------------------------------------
536    Name: Matthias Blume (blume (at) tti - c (dot) org)
537    Date: 2004/12/13 14:45:00 CST
538    Tag: blume-20041213-ml-makedepend
539    Description:
540    
541    Fixed bug in code for ml-makedepend.
542    
543    ----------------------------------------------------------------------
544    Name: Matthias Blume (blume (at) tti - c (dot) org)
545    Date: 2004/12/09 16:30:00 CST
546    Tag: blume-20041209-statistics
547    Description:
548    
549    Added two simple but potentially useful statistics modules to SML/NJ Library.
550    (See CHANGES file there.)
551    
552    ----------------------------------------------------------------------
553    Name: Matthias Blume (blume (at) tti - c (dot) org)
554    Date: 2004/12/01 16:50:00 CST
555    Tag: blume-20041201-atom
556    Description:
557    
558    smlnj-lib:
559    
560      Added function HashString.hashString' for substrings.
561      Hand-inlined CharVector.foldl into HashString (for speed).
562      Modified implementation of structure Atom to avoid extracting
563      strings from substrings unless necessary.
564      (Also see CHANGES file for smlnj-lib.)
565    
566    ----------------------------------------------------------------------
567    Name: Matthias Blume (blume (at) tti - c (dot) org)
568    Date: 2004/11/24 22:15:00 CST
569    Tag: blume-20041124-cml
570    Description:
571    
572    Made sure CML compiles when Position = Int64.
573    
574    ----------------------------------------------------------------------
575    Name: Matthias Blume (blume (at) tti - c (dot) org)
576    Date: 2004/11/24 14:30:00 CST
577    Tag: blume-20041124-position
578    Description:
579    
580    The compiler can now be compiled in a mode that makes structure Position
581    equal to Int64.  The default, however, is unchanged (Position = Int31)
582    for the time being.
583    
584    To enable 64-bit positions, use the following procedure:
585    
586    1. Start sml
587    2. Autoload $smlnj/cmb.cm (if not already autoloaded)
588    3. Type
589          #set (CMB.symval "USE_64_BIT_POSITIONS") (SOME 1);
590    4. Run CMB.make() as usual.
591    
592    This is barely tested.  The only test so far was a little SML program
593    counting the number of characters in an 8-gigabyte file by
594    reading it character-by-character.  That test was successful.
595    
596    In support of 64-bit positions, a number of new functions have been
597    added to the runtime system.
598    
599    ----------------------------------------------------------------------
600    Name: Matthias Blume (blume (at) tti - c (dot) org)
601    Date: 2004/11/23 14:45:00 CST
602    Tag: blume-20041123-useFile
603    Description:
604    
605    Fixed a problem with unhelpful error messages related to problems with
606    .cm- or .sml files that appear as part of the sml command line.
607    
608    ----------------------------------------------------------------------
609    Name: Matthias Blume (blume (at) tti - c (dot) org)
610    Date: 2004/11/18 15:40:00 CST
611    Tag: Release_110_51
612    Description:
613    
614    New working version (110.51).  NEW BOOTFILES!
615    
616    ----------------------------------------------------------------------
617    Name: Matthias Blume (blume (at) tti - c (dot) org)
618    Date: 2004/11/18 15:35:00 CST
619    Tag: <none>
620    Description:
621    
622    Enabled dlopen and friends for FreeBSD (as recommended by Johannes 5
623    Joemann).
624    
625    ----------------------------------------------------------------------
626    Name: Allen Leung (leunga (at) reservoir (dot) com)
627    Date: 2004/11/17 16:05:21 EST 2004
628    Tag: leunga-20041117-mlrisc-live-kill
629    Description:
630    
631       Added support for MLTree constructs LIVE and KILL
632    to all the architectures.
633    
634    ----------------------------------------------------------------------
635    Name: Matthias Blume (blume (at) tti - c (dot) org)
636    Date: 2004/11/13 00:20:00 CST
637    Tag: blume-20041113-versiontool
638    Description:
639    
640    - Stripped down the versiontool: It now only handles the version number.
641      The date string is generated at bootstrap time (during makeml).
642    
643    - In a previous commit, fixed a minor issue with how polyequal is being
644      translated.  In particular, the code now "looks through" abstractions.
645      This results in slightly fewer polyEqual warnings and hopefully slightly
646      more efficient code.  Important examples for where this matters are
647      the new int64 and word64 types.
648    
649    ----------------------------------------------------------------------
650    Name: Matthias Blume (blume (at) tti - c (dot) org)
651    Date: 2004/11/12 00:30:00 CST
652    Tag: blume-20041112-int64
653    Description:
654    
655    Structure Int64 fully hooked in.  (The implementation is not very
656    efficient, though.)
657    
658    ----------------------------------------------------------------------
659    Name: Matthias Blume (blume (at) tti - c (dot) org)
660    Date: 2004/11/11 17:30:00 CST
661    Tag: blume-20041111-more64
662    Description:
663    
664    All the pieces of Word64 are now there, with the exception of the
665    conversions from and to LargeWord. (Eventually these need to be identities,
666    but for the time being they don't even make sense because LargeWord is
667    32-bit wide.)
668    
669    Also started to add similar support for Int64, but major pieces of that
670    are still missing.
671    
672    ----------------------------------------------------------------------
673    Name: Matthias Blume (blume (at) tti - c (dot) org)
674    Date: 2004/11/11 00:15:00 CST
675    Tag: blume-20041111-word64
676    Description:
677    
678    Structure Word64 is now (almost) complete, word literals and patterns
679    seem to work.  There are a few odd pieces missing.  In particular,
680    I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
681    at the moment.
682    
683    Making Word64 official would mean that LargeWord becomes Word64.  But
684    this requires extreme care because most word-word conversions have to
685    go through LargeWord, so making a mistake means loss of efficiency or
686    worse.  Eventually there will be a solution similar to (but actually
687    simpler than) what I did with IntInf.
688    
689    ----------------------------------------------------------------------
690    Name: Matthias Blume (blume (at) tti - c (dot) org)
691    Date: 2004/11/10 18:12:00 CST
692    Tag: blume-20041110-64bit
693    Description:
694    
695    More 64-bit hacking (but still not even half-way there yet).
696    Also, some assorted improvements to the handling of 8-bit words.
697    
698    ----------------------------------------------------------------------
699    Name: Matthias Blume (blume (at) tti - c (dot) org)
700    Date: 2004/11/09 17:50:00 CST
701    Tag: <none>
702    Description:
703    
704    Started adding some infrastructure for supporting 64-bit int- and
705    word-types.  (Still in its very early stages.)
706    
707    ----------------------------------------------------------------------
708    Name: Matthias Blume (blume (at) tti - c (dot) org)
709    Date: 2004/10/28 10:45:00 CDT
710    Tag: Release_110_50
711    Description:
712    
713    New working version (110.50).  NEW BOOTFILES!
714    
715                 =====================
716    
717    Also:
718    
719      - Changend config/srcarchiveurl from a file just
720        containing the URL string into a file containing
721        shell script code.  The code has access to the $VERSION variable.
722      - Made corresponding changes to config/install.sh and config/unpack.
723      - Default contents of config/srcarchiveurl uses $VERSION and
724        normally does not have to be edited to reflect a version change.
725    
726      (As a result, a version change can be done by just editing
727       config/version, the rest is now automatic.)
728    
729    ----------------------------------------------------------------------
730    Name: Matthias Blume (blume (at) tti - c (dot) org)
731    Date: 2004/10/27 17:50:00 CDT
732    Tag: blume-20041027-btrace-msg
733    Description:
734    
735    BackTrace.monitor now also reports the source of the exception that
736    triggered the trace.
737    
738    ----------------------------------------------------------------------
739    Name: Matthias Blume (blume (at) tti - c (dot) org)
740    Date: 2004/10/27 17:20:00 CDT
741    Tag: blume-20041027-x86-c-calls
742    Description:
743    
744    This is the HISTORY entry for two earlier commits, both concerning
745    the x86 c-calls code in MLRISC:
746    
747      - added a missing LOAD in the code that deals with struct arguments
748      - made sure the caller does not add the wrong number of bytes to the
749        stack pointer after a call of a function returning a struct
750        (the callee already pops the implicit argument which points to
751         the space reserved for the result)
752    
753    ----------------------------------------------------------------------
754    Name: Allen Leung (leunga (at) reservoir (dot) com)
755    Date: 2004/10/24 14:00:00 EST
756    Tag: leunga-20041024-x86-gas-fucomip
757    Description:
758    
759       John discovered a bug in the syntax of fucomip.
760       The opcodes FU?COMIP? have been changed to
761    
762           fu?comip? %st(i), %st
763    
764    ----------------------------------------------------------------------
765    Name: Matthias Blume (blume (at) tti - c (dot) org)
766    Date: 2004/10/20 15:06:00 CDT
767    Tag: blume-20041020-standalone-backtrace
768    Description:
769    
770    Added a mechanism for getting back-trace information from standalone
771    programs.  Here is how it works:
772    
773       1. The part of the program from which you want to get backtrace
774          information (usually the whole program) should be wrapped with
775          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
776          main program could be modified from something like
777    
778             fun main (pgm, args) = ...
779    
780         to
781    
782            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
783    
784       2. To be able to access BackTrace.monitor, you have to add
785          library $smlnj-tdp/plugins.cm to the .cm file that contains your
786          main function.
787    
788       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
789          CM might have created in the past for your project).
790    
791       4. Build the system using this command line:
792    
793             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
794                      myprog.cm MyProg.main myprog
795    
796          instead of the usual
797    
798             ml-build myprog.cm MyProg.main myprog
799    
800    I changed a library name:
801    
802       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
803    
804    New libraries:
805    
806       $smlnj-tdp/back-trace.cm
807          -- when loaded causes the back-trace plugin to be installed
808       $smlnj-tdp/coverage.cm
809          -- when loaded causes the coverage plugin to be installed
810    
811    ----------------------------------------------------------------------
812    Name: Matthias Blume (blume (at) tti - c (dot) org)
813    Date: 2004/10/18 16:45:00 CDT
814    Tag: blume-20041018-groupowner
815    Description:
816    
817    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
818    
819    Eliminated group owner specs from .cm files throughout the source tree.
820    
821    ----------------------------------------------------------------------
822    Name: Matthias Blume (blume (at) tti - c (dot) org)
823    Date: 2004/10/15 15:45:00 CDT
824    Tag: blume-20041015-coverage
825    Description:
826    
827    * Test coverage tool added!
828    * Further reorganization of tracing-, debugging-, and profiling support:
829    
830        - moved original BTImp -- now called BackTrace -- into a separate
831          library called $/trace-debug-profile.cm
832        - eliminated all mentions of BTrace from SMLofNJ.Internals
833        - only the instrumentation mechanism is now left in the compiler proper
834        - BackTrace module is a plugin which is NOT plugged in by default
835        - Coverage module is another such plugin
836    
837    To get the benefits of any of these plugin modules, the code in
838    question must be compiled with tdp instrumentation turned on.  This
839    can be done by setting SMLofNJ.Internals.TDP.mode to true.
840    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
841    
842    Plugins are selected at link time.  (Pre-compiled instrumented code
843    can be re-loaded with different plugins in effect.)  When an
844    instrumented module is linked, whatever plugins are at that time
845    enabled will come into effect for that module.
846    
847    To enable the back-trace plugin, load library $/trace-debug-profile.cm
848    and invoke BackTrace.install() (e.g., from the interactive prompt).
849    To enable the coverage plugin, load the same library and invoke
850    Coverage.install().
851    
852    Back-traces are generated automatically on uncaught exceptions and
853    when the code in question explicitly invokes BackTrace.trigger().
854    
855    Coverage (and execution frequency-) information must be queried
856    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
857    
858    ----------------------------------------------------------------------
859    Name: Matthias Blume (blume (at) tti - c (dot) org)
860    Date: 2004/10/14 17:40:00 CDT
861    Tag: blume-20041014-tdp-core
862    Description:
863    
864    Snapshot of a significant overhaul of how the trace/debug/profile support
865    is hooked into the system (specifically: Core and SMLofNJ.Internals).
866    
867    ----------------------------------------------------------------------
868    Name: Matthias Blume (blume (at) tti - c (dot) org)
869    Date: 2004/10/13 16:34:00 CDT
870    Tag: blume-20041013-tdp
871    Description:
872    
873    Some rationalization of names:
874    
875       structure BTrace -> structure TDPInstrument
876       etc.
877    
878    This is is preparation of using the original back-trace
879    instrumentation for other purposes.  "TDP" stands for
880    Trace/Debug/Profile.
881    
882    The control flag controlling whether instrumentation is on or off is now
883    registered under a different name, so instead of running sml as
884    
885     sml -Cinstrument.btrace-mode=true
886    
887    one has to say
888    
889     sml -Ctdp.instrument=true
890    
891    ----------------------------------------------------------------------
892    Name: Matthias Blume (blume (at) tti - c (dot) org)
893    Date: 2004/10/11 16:37:00 CDT
894    Tag: blume-20041011-regions
895    Description:
896    
897    Made some minor modifications to elabcore.sml to have source regions
898    be propagated more tightly -- resulting in better (i.e., smaller)
899    regions being reported in error- and debug messages.
900    
901    ----------------------------------------------------------------------
902    Name: Matthias Blume (blume (at) tti - c (dot) org)
903    Date: 2004/10/08 22:50:00 CDT
904    Tag: blume-20041008-cmkw
905    Description:
906    
907    Fixed handling of keywords in .cm files: After seeing "is" the lexer
908    treats subsequent occurrences of "group", "library", "source", "is",
909    "*", and "-" as ordinary identifiers rather than keywords.
910    
911    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
912    is supposed to accept a tool argument called "source", but this did
913    not work because of the clash with the keyword.
914    
915    ----------------------------------------------------------------------
916    Name: Matthias Blume (blume (at) tti - c (dot) org)
917    Date: 2004/10/07 16:00:00 CDT
918    Tag: blume-20041007-cleanup
919    Description:
920    
921    Assorted cleanup work:
922    
923      - got rid of intstrmap in favor of using the library's
924        hash table implementation
925      - threw out most of the pathnames stuff, as it was not used anyway
926      - simplified tokentable implementation
927      - fixed some minor spelling errors
928    
929    ----------------------------------------------------------------------
930    Name: Matthias Blume (blume (at) tti - c (dot) org)
931    Date: 2004/10/06 15:15:15 CDT
932    Tag: blume-20041006-handler
933    Description:
934    
935    Cleaned up the absyn to reflect the invariant that HANDLE always
936    carries a FNexp as part of the type definition.  This eliminates some
937    superfluous sanity checks at runtime down the road.
938    
939    Some minor cleanup of the btrace code.
940    
941    ----------------------------------------------------------------------
942    Name: Matthias Blume (blume (at) tti - c (dot) org)
943    Date: 2004/10/01 10:20:30 CDT
944    Tag: blume-20041001-slave
945    Description:
946    
947    Added hack to make slave mode work in the presence of the version
948    tool.  (Still, since the master does two passes over the code for
949    CMB.make, the release number gets bumped twice when slaves are
950    attached.  I don't know if this is worth fixing...)
951    
952    ----------------------------------------------------------------------
953    Name: Matthias Blume (blume (at) tti - c (dot) org)
954    Date: 2004/09/30 10:55:00 CDT
955    Tag: blume-20040930-version
956    Description:
957    
958    * Moved the "version" magic into its own little library under
959      src/system/smlnj/internal.  This avoids expensive reconstruction of
960      a stable src/compiler/core.cm.
961    
962    * At the same time, structure CompilerVersion is now known as
963      structure SMLNJVersion.
964    
965    * Arranged for the version tool to NOT kick in when rebuilding the system
966      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
967      Also, loading the versiontool does not work when rebuilding the system
968      because CM is not properly initialized at that time.
969    
970    ----------------------------------------------------------------------
971    Name: Matthias Blume (blume (at) tti - c (dot) org)
972    Date: 2004/09/29 14:00:00 CDT
973    Tag: blume-20040929-autoversion
974    Description:
975    
976    Implemented some CM magic to have
977        file src/compiler/TopLevel/main/version.sml
978    generated automagically.
979    The version is taken from two files: config/version and config/release.
980    The first is expected to contain a two-part version number such as 110.49.
981    The second should contain a single number, but it may be missing.
982    
983    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
984    time the version tool is loaded (which is the first time you say CMB.make),
985    then the tool will increment the value stored in config/release every
986    time CMB.make is invoked.
987    
988    The binfile format is now insensitive to anything beyond the first
989    two components of a version number, so bumping the release does not render
990    binfiles incompatible.  Auto-bumping can be used to keep track of versions
991    during development without invalidating existing binfiles.
992    
993    In any case, every CMB.make updates the date information in version.sml.
994    (This is the date that is printed in the banner.)
995    
996    ----------------------------------------------------------------------
997    Name: Matthias Blume (blume (at) tti - c (dot) org)
998    Date: 2004/09/28 10:53:00 CDT
999    Tag: blume-20040928-controls
1000    Description:
1001    
1002    Some cleanup of the controls code.
1003    
1004    ----------------------------------------------------------------------
1005    Name: Matthias Blume (blume (at) tti - c (dot) org)
1006    Date: 2004/09/27 22:08:00 CDT
1007    Tag: blume-20040927-controls
1008    Description:
1009    
1010    Added two pieces of functionality to the Controls interface:
1011    
1012    1.   val save'restore: 'a control -> unit -> unit
1013    
1014       grabs the current value of the control in stage 1 and restores it
1015       in stage 2.
1016    
1017    2.   val set' : 'a control * 'a -> unit -> unit
1018    
1019       stores the given value into the control in stage 2 (i.e., delayed)
1020       but does all error checking in stage 1.
1021       (This is for string controls that need to do parse their argument --
1022        something that might fail.  In some cases, notably in CM, one
1023        already knows the intended argument but wants to delay the actual
1024        assignment until a time when error recovery would be more difficult.)
1025    
1026    Changed the handling of controls in tool arguments to classes "sml" and
1027    "lazysml":
1028       - use Controls.save'restore as a more robust way of restoring the
1029         old value (in particular: without having to re-parse the string)
1030       - use controls to handle the "overload" keyword in the init group
1031         (I believe this change actually fixes a long-standing obscure bug.)
1032    
1033    ----------------------------------------------------------------------
1034    Name: Matthias Blume (blume (at) tti - c (dot) org)
1035    Date: 2004/09/27 17:00:00 CDT
1036    Tag: blume-20040927-lazysml
1037    Description:
1038    
1039    Added a new tool class called "lazysml" to CM's tool chest.  The only
1040    difference to "sml" is that compilation is done with Control.lazysml
1041    set to true.  A source of class "lazysml" is automatically recognized
1042    by a file name suffix of ".lml".
1043    
1044    In addition to the above feature, the original class "sml" now also
1045    supports a tool argument "lazy" which has the same effect.  As a
1046    result, the following three lines are equivalent:
1047    
1048        foo.sml : lazysml
1049        foo.sml : sml (lazy)
1050        foo.sml (lazy)
1051    
1052    The setting goes into effect both during parsing and during
1053    compilation.  The original setting is restored right after parsing and
1054    after compilation, respectively.
1055    
1056    In addition to all the above, there is also a general mechanism to set
1057    ANY of the "controls" that are available at the command line via
1058    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
1059    apply as well. (In fact, "lazy" is implemented as a special case of
1060    the general mechanism.)
1061    
1062    The .cm file syntax uses a new keyword tool argument called "with".
1063    There are several ways of indicating the desired settings:
1064    
1065        foo.sml (with:parser.quotations=true)
1066        foo.sml (with:(name:parser.quotations value:true))
1067        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
1068        foo.sml (with:(name1=value1 name2=value2 ...))
1069        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
1070    
1071        etc.
1072    
1073    Another possible abbreviation is to leave out the =v or value:v part
1074    if the name refers to a boolean control (in which case the value is
1075    taken to be true).  Thus, one could get lazy sml also by saying:
1076    
1077        foo.sml (with:parser.lazy-keyword=true)
1078        foo.sml (with:parser.lazy-keyword)
1079        foo.sml (with:(name:parser.lazy-keyword value:true))
1080        foo.sml (with:(name:parser.lazy-keyword))
1081    
1082    ----------------------------------------------------------------------
1083    Name: Matthias Blume (blume (at) tti - c (dot) org)
1084    Date: 2004/09/24 16:22:00 CDT
1085    Tag: blume-20040924-ppc-long-branch
1086    Description:
1087    
1088    Turned message about "emiting long form of branch" off by default.
1089    Added a control flag to turn it back on when desired.
1090    
1091    ----------------------------------------------------------------------
1092    Name: Matthias Blume (blume (at) tti - c (dot) org)
1093    Date: 2004/09/24 16:05:00 CDT
1094    Tag: blume-20040924-rounding
1095    Description:
1096    
1097    Applied patch for setting rounding modes under Mac OS X.  Thanks to
1098    Melissa O'Neill for providing the code!
1099    
1100    ----------------------------------------------------------------------
1101    Name: Matthias Blume (blume (at) tti - c (dot) org)
1102    Date: 2004/09/23 17:30:00 CDT
1103    Tag: blume-20040923-envvars
1104    Description:
1105    
1106    1. Changed definition of type ControlRegistry.registry_tree to
1107       include control_info (i.e., the name of the controlling environment
1108       variable).
1109    
1110    2. Added command-line flags -e and -E to print the names of environment
1111       variables that can be used to control internal settings.
1112       (This uses the new API mentioned in 1.)
1113    
1114    ----------------------------------------------------------------------
1115    Name: Matthias Blume (blume (at) tti - c (dot) org)
1116    Date: 2004/09/13 16:50:00 CDT
1117    Tag: Release_110_49
1118    Description:
1119    
1120    New working version (110.49).  NEW BOOTFILES!
1121    
1122    ----------------------------------------------------------------------
1123    Name: Matthias Blume
1124    Date: 2004/09/13 16:20:00 CDT
1125    Tag: blume-20040913-config-mlrisc
1126    Description:
1127    
1128    Put target "mlrisc" back into the default list.
1129    (There is no harm in having it, and some users have expressed their
1130    wish to have "mlrisc" included by default.)
1131    
1132    ----------------------------------------------------------------------
1133    Name: John Reppy
1134    Date: 2004/09/13
1135    Tag: jhr-20040913-signals
1136    Description:
1137    
1138    Fixed the signal masking code to properly nest mask/unmask operations
1139    on a per-signal basis.
1140    
1141    ----------------------------------------------------------------------
1142    Name: Matthias Blume (blume (at) tti - c (dot) org)
1143    Date: 2004/09/08 13:20:00 CDT
1144    Tag: blume-20040908-heap-magic
1145    Description:
1146    
1147    Bumped the heap macig number to 0x09082004 to account for the changed
1148    layout of the ML frame under MacOS X.
1149    
1150    ----------------------------------------------------------------------
1151    Name: Allen Leung (leunga (at) reservoir (dot) com)
1152    Date: 2004/09/03 11:26:00 EST
1153    Tag: leunga-20040903-cygwin-install
1154    Description:
1155    
1156       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
1157    The Cygwin runtime is turned on by setting the environment
1158    variable SMLNJ_CYGWIN_RUNTIME to 1.
1159    
1160    ----------------------------------------------------------------------
1161    Name: Matthias Blume (blume (at) tti - c (dot) org)
1162    Date: 2004/08/31 17:15:00 CDT
1163    Tag: blume-20040831-core
1164    Description:
1165    
1166    Added some exports to src/compiler/core.cm upon request by J. Joemann.
1167    
1168    ----------------------------------------------------------------------
1169    Name: Matthias Blume (blume (at) tti - c (dot) org)
1170    Date: 2004/08/30 17:55:00 CDT
1171    Tag: blume-20040830-installer
1172    Description:
1173    
1174    Upon request by Johannes Joemann:
1175    
1176    - improved ML code of installer to fall back to coping when renaming
1177      fails (i.e., when source and target are on different file systems);
1178      the code compiles but has yet to be tested in anger
1179    - removed mlrisc from list of default targets (config/targets)
1180    
1181    ----------------------------------------------------------------------
1182    Name: Matthias Blume (blume (at) tti - c (dot) org)
1183    Date: 2004/08/27 17:20:00 CDT
1184    Tag: blume-20040827-ptreql
1185    Description:
1186    
1187    Added ptreql primop to structure InlineT (upon request from Larry
1188    Paulson).
1189    
1190    ----------------------------------------------------------------------
1191    Name: Allen Leung (leunga (at) reservoir (dot) com)
1192    Date: 2004/08/15 21:21:00 EST
1193    Tag: leunga-110_48-udgraph
1194    Description:
1195    
1196    Another bug fix from Carl Hauser:
1197    
1198    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
1199    > 48c48
1200    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
1201    > e::L)
1202    > ---
1203    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
1204    > Without this, any deletion of an edge in an undirected graph does severe
1205    > violence to the graph.
1206    
1207    ----------------------------------------------------------------------
1208    Name: Allen Leung (leunga (at) reservoir (dot) com)
1209    Date: 2004/08/10 23:35:00 EST
1210    Tag: leunga-110_48-ppc
1211    Description:
1212    
1213       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
1214    
1215    ----------------------------------------------------------------------
1216    Name: Matthias Blume (blume (at) tti - c (dot) org)
1217    Date: 2004/08/10 12:00:00 CDT
1218    Tag: Release_110_48
1219    Description:
1220    
1221    New working version (110.48).  NEW BOOTFILES!
1222    
1223    ----------------------------------------------------------------------
1224    Name: Allen Leung (leunga (at) reservoir (dot) com)
1225    Date: 2004/08/09 12:21:00 EST
1226    Tag: leunga-110_47-dijsktra
1227    Description:
1228    
1229    Bug fix from Carl Hauser:
1230    
1231    single_source_shortest_paths in dijkstra.sml was observed to get wrong
1232    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
1233    
1234    The problem is that following the expression A.update(dist,s,Num.zero)
1235    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
1236    
1237    ----------------------------------------------------------------------
1238    Name: Matthias Blume (blume (at) tti - c (dot) org)
1239    Date: 2004/08/06 18:10:00 CDT
1240    Tag: blume-20040806-cmdline
1241    Description:
1242    
1243    Fiddled with handling of command-line options:
1244    
1245      * sml now quits after processing the command line
1246        if -H, -S, -h<n>, or -s<n> appears as the last
1247        command-line argument
1248      * a new option -q terminates the session when encountered on
1249        the command line; subsequent arguments will be ignored
1250      * bug fixes: short (erroneous) arguments are no longer ignored
1251        completely
1252    
1253    ----------------------------------------------------------------------
1254    Name: Allen Leung (leunga (at) reservoir (dot) com)
1255    Date: 2004/08/04 18:17:00 EST
1256    Tag: leunga-110_47-ppc-ibm-asm
1257    Description:
1258    
1259      - Added minimal IBM assembly syntax support for PowerPC.
1260    
1261      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
1262        been altered in the runtime.  We need an automatic way to keep the file
1263        in sync.
1264    
1265    ----------------------------------------------------------------------
1266    Name: Matthias Blume (blume (at) tti - c (dot) org)
1267    Date: 2004/08/04 14:00:00 CDT
1268    Tag: Release_110_47
1269    Description:
1270    
1271    New working version (110.47).  NEW BOOTFILES!
1272    
1273    ----------------------------------------------------------------------
1274    Name: Matthias Blume (blume (at) tti - c (dot) org)
1275    Date: 2004/08/03 14:25:00 CDT
1276    Tag: blume-20040803-callingconv
1277    Description:
1278    
1279    Added low-level support for choosing C calling conventions by
1280    twiddling the type of rawccall.  (See
1281    src/compiler/Semant/types/cproto.sml for details.)
1282    
1283    ----------------------------------------------------------------------
1284    Name: Matthias Blume (blume (at) tti - c (dot) org)
1285    Date: 2004/08/02 15:55:00 CDT
1286    Tag: blume-20040802-backout
1287    Description:
1288    
1289    Backed out of change to win32-filesys.c.  The earlier patch to
1290    get_file_time caused CM to produce files with the wrong time stamp.
1291    
1292    ----------------------------------------------------------------------
1293    Name: Matthias Blume (blume (at) tti - c (dot) org)
1294    Date: 2004/08/02 14:45:00 CDT
1295    Tag: blume-20040802-nlffi-win32
1296    Description:
1297    
1298    Added NLFFI support for Win32, adapted from a patch provided by David
1299    Hansel.  This is currently completely untested.  Also, the issue
1300    concerning stdcall vs. ccall is still unresolved.
1301    
1302    ----------------------------------------------------------------------
1303    Name: Matthias Blume (blume (at) tti - c (dot) org)
1304    Date: 2004/07/30 17:55:00 CDT
1305    Tag: blume-20040730-various
1306    Description:
1307    
1308    Gearing up towards 110.47...
1309    
1310    - various minor bugfixes to ml-nlffigen
1311    - a beginning of a manual for nlffi
1312    
1313    - eliminated 'export name=value' in config/install.sh as this does
1314      not work with certain versions of /bin/sh
1315      (Thanks to David King at Motorola for catching this.)
1316    
1317    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
1318      - added a test for tm==NULL to gmtime.c and localtime.c
1319      - applied patch for incorrect GetFileTime under win32
1320      - toSeconds -> toMilliseconds in Win32/win32-process.sml
1321    
1322    ----------------------------------------------------------------------
1323    Name: Matthias Blume (blume (at) tti - c (dot) org)
1324    Date: 2004/07/21 18:20:00 CDT
1325    Tag: blume-20040721-nlffigen
1326    Description:
1327    
1328    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
1329      for a typedef to an incomplete type, but leave out the "typ" member.
1330      (This is just for consistency.)
1331    - Started to produce what is supposed to become better (i.e., comprehensive)
1332      documentation of what ml-nlffigen does and produces.
1333    
1334    ----------------------------------------------------------------------
1335    Name: Matthias Blume (blume (at) tti - c (dot) org)
1336    Date: 2004/07/14 16:25:00 CDT
1337    Tag: blume-20040714-union
1338    Description:
1339    
1340    Added C_UNION to c-calls/c-types.sml and updated the machinery
1341    (ml-nlffigen, cproto.sml) that conveys C function interface
1342    information to the code generator.
1343    
1344    However, the actual architecture-specific implementation of function
1345    arguments and results that are C unions is still not implemented.
1346    
1347    ----------------------------------------------------------------------
1348    Name: Allen Leung (leunga (at) reservoir (dot) com)
1349    Date: 2004/07/14 14:38:00 EST
1350    Tag: leunga-110_46_1-ppc-lwzu
1351    Description:
1352    
1353        Added these instructions to the PowerPC architecture:
1354        LBZU(X), LHZU(X), LWZU(X),
1355        STWU(X), STFDU, STFSU
1356    
1357        etc...
1358    
1359        Note: I haven't added their instruction encoding into the description.
1360    
1361    ----------------------------------------------------------------------
1362    Name: Allen Leung (leunga (at) reservoir (dot) com)
1363    Date: 2004/07/13 15:04:00 EST
1364    Tag: leunga-110_46_1-ppc-lwarx
1365    Description:
1366    
1367       Added the two instructions LWARX and STWCX to the PowerPC
1368    instruction set.
1369    
1370       A (untested) rewrite of loop-structure.sml.  The old version
1371    is completely broken.
1372    
1373    ----------------------------------------------------------------------
1374    Name: Matthias Blume (blume (at) tti - c (dot) org)
1375    Date: 2004/07/13 13:50:00 CDT
1376    Tag: blume-20040713-nlffi
1377    Description:
1378    
1379    - use paramAlloc to report c-calls with too many arguments
1380      (for PPC version where parameter area is pre-allocated)
1381    - added ccall_maxargspace to machspec (to implement the above)
1382    - made "make" commend in CM's "make" tool configurable
1383    - added option (default: on) for passing the name of the SML/NJ's "bin"
1384      directory to "make"; the call looks like this:
1385    
1386         make <options> SMLNJ_BINDIR=<dir> <target>
1387    
1388      This can be used by the Makefile to, e.g., pick the "right" version
1389      of ml-nlffigen.
1390    - minor code tweaks
1391    
1392    ----------------------------------------------------------------------
1393    Name: Matthias Blume (blume (at) tti - c (dot) org)
1394    Date: 2004/07/12 22:50:00 CDT
1395    Tag: blume-110_46_1-macosx-nlffi
1396    Description:
1397    
1398    NLFFI under Mac OS X now working (sort of).  This is largely untested,
1399    though.
1400    
1401    Note:  1. You have to make a new, clean build of the runtime system.
1402           2. There are new BOOTFILES, you have to use them!
1403              (Doing the bootstrap process yourself would be *very* painful!
1404               If you absolutely have to do it, build the system under
1405               a different architecture and then cross-compile.)
1406    
1407    Version bumped to 110.46.1 to account for runtime data format changes.
1408    
1409    ----------------------------------------------------------------------
1410    Name: Matthias Blume (blume (at) tti - c (dot) org)
1411    Date: 2004/06/18 14:30:00 CDT
1412    Tag: blume-20040618-unix
1413    Description:
1414    
1415    Changed the implementation of structure Unix so that the same stream
1416    is returned every time one of the {text,bin}{In,Out}streamOf functions
1417    is invoked on the same proc.  This is not what the spec currently
1418    says -- although IMO it arguably should.  (See discussion below.)
1419    
1420    ----------------------------------------------------------------------
1421    Name: Matthias Blume (blume (at) tti - c (dot) org)
1422    Date: 2004/06/17 18:15:00 CDT
1423    Tag: Release_110_46
1424    Description:
1425    
1426    New working version (110.46).  NEW BOOTFILES!
1427    
1428    ----------------------------------------------------------------------
1429    Name: Matthias Blume (blume (at) tti - c (dot) org)
1430    Date: 2004/06/17 17:20:00 CDT
1431    Tag: blume-20040617-timer-unix
1432    Description:
1433    
1434    Changed the interface of structures Timer and Unix to match the most
1435    recent Basis spec.
1436    
1437    In the case of Unix there still seems to be an open/weird issue:
1438    
1439       The {text,bin}{In,Out}streamOf functions are supposed to create
1440       fresh streams whenever they are called -- as opposed to have them
1441       return the same stream every time.  This design is supposed to
1442       prevent space leaks caused by proc values hanging on to streams.
1443    
1444       The reap function, on the other hand, is supposed to close the
1445       streams.  This cannot be done without having a handle on the
1446       stream in proc after all...
1447    
1448    I took the liberty to implement the following stopgap solution:
1449    
1450      The proc value hangs on to the most recently created stream(s).
1451      Reap closes those.  If either or both of the two streams hadn't
1452      been created at all yet, then reap will close the corresponding
1453      file descriptors directly.
1454    
1455    PS: I don't understand the original space leak argument anymore.  If
1456    a proc hangs on to the imperative stream, then I/O operations on those
1457    will advance the state of the cached stream and avoid the space leak.
1458    
1459    ----------------------------------------------------------------------
1460    Name: Matthias Blume (blume (at) tti - c (dot) org)
1461    Date: 2004/05/28 16:45:00 CDT
1462    Tag: blume-20040528-basis
1463    Description:
1464    
1465    Added signature PACK_REAL and exported functor PrimIO.
1466    
1467    ----------------------------------------------------------------------
1468    Name: Matthias Blume (blume (at) tti - c (dot) org)
1469    Date: 2004/05/25 16:00:00 CDT
1470    Tag: blume-20040525-group-owner
1471    Description:
1472    
1473    CM now ignores (but still accepts) the "owner" information in group
1474    descriptions.  The owner of a group is its next enclosing
1475    library. Each group must have a unique owner.  (There is a virtual
1476    "toplevel" library that owns groups which are not nested within a real
1477    library.)  Previously, each group had to explicitly declare its owner,
1478    and CM would check that such a declaration is correct.  The new scheme
1479    is to have CM check that for each group there is precisely one owning
1480    library.
1481    
1482    The advantage of the new scheme is that the programmer no longer needs
1483    to maintain the somewhat annoying owner information.  The downside is
1484    that CM cannot enforce the ownership rule across multiple runs of
1485    CM.make.  Fortunately, enclosing the same group in two different
1486    libraries A and B which are not part of the same program does not
1487    cause real problems.
1488    
1489    ----------------------------------------------------------------------
1490    Name: Matthias Blume (blume (at) tti - c (dot) org)
1491    Date: 2004/05/20 16:00:00 CDT
1492    Tag: blume-20040520-win32
1493    Description:
1494    
1495    Made the win32 version work again.  (Strangely, a misplaced comma had
1496    slipped into win32-process.c which prevented the runtime from being
1497    compiled correctly.)
1498    
1499    Also, included a minor addition to ml-build.bat analogous to what was
1500    done in blume-20040519-ml-build.
1501    
1502    ----------------------------------------------------------------------
1503    Name: Matthias Blume (blume (at) tti - c (dot) org)
1504    Date: 2004/05/19 22:10:00 CDT
1505    Tag: blume-20040519-ml-build
1506    Description:
1507    
1508    Arranged for ml-build to clean up after itself a little bit better.
1509    The script generates a temporary SML source file and compiles it using
1510    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
1511    gets rid of those at the end, so they don't accumulate under .cm.
1512    
1513    This required a minor change to install.sh because the name of the
1514    metadata directory (default: .cm) is actually configurable at
1515    installation time.
1516    
1517    ----------------------------------------------------------------------
1518    Name: Matthias Blume (blume (at) tti - c (dot) org)
1519    Date: 2004/05/18 15:50:00 CDT
1520    Tag: blume-20040518-mkreader
1521    Description:
1522    
1523    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
1524    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
1525    
1526    ----------------------------------------------------------------------
1527    Name: Matthias Blume (blume (at) tti - c (dot) org)
1528    Date: 2004/05/11 14:35:00 CDT
1529    Tag: blume-20040511-win32sock
1530    Description:
1531    
1532    Added previously missing support for many socket-related functions
1533    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
1534    for the voluminous patch!
1535    
1536    (I have not tested this patch under win32 yet.)
1537    
1538    Here is David's e-mail:
1539    
1540    Hi,
1541    
1542    Attached to this email you find a diff against sml/nj 110.45
1543    that will enable socket support under Windows.
1544    
1545    To apply the patch (using unix or cygwin)
1546    1) gunzip runtime.diff.gz
1547    2) "cd" into "src/runtime" in the source tree of a fresh
1548       110.45 installation.
1549    3) patch -p 1 < [your/path/to]runtime.diff
1550    
1551    The code compiles fine but has NOT yet been extensively tested.
1552    I only ran a few tests for basic socket client functionality
1553    (which worked fine).  Especially the functions that use ioctl
1554    are not tested at all and might not work (see below).
1555    
1556    I implemented this since we want to move to a newer version of sml/nj
1557    but need socket support in order to use it.  This is the first time I
1558    even had a look at the sml/nj source,  so please review my changes
1559    before making this part of the distribution!  Here are a few issues
1560    that I think might be better for someone to solve who is more
1561    familiar with the sml/nj source (and socket programming):
1562    
1563    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
1564      exception since I could not figure out what the windows equivalent
1565      of these functions is
1566    
1567    - In sockets-osdep.h there are a some #include statements that are
1568      only used in a few files that include sockets-osdep.h
1569    
1570    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
1571      does not process its return value since I don't know how to
1572      report an error upwards.
1573    
1574    - It would probably be good to have a call to WSACleanup() when
1575      the library is unloaded (if there is such a possibility).
1576      Otherwise I think Windows will take care of this automatically
1577      when the process finishes.
1578    
1579    - I used ioctlsocket() as a replacement for ioctl() but I have
1580      no idea if that is actually the proper replacement on Windows.
1581    
1582    - All these issues are marked in the code by "FIXME" comments.
1583    
1584    We use sml/nj extensively in our products and are quite happy
1585    with it.  I hope this contribution will help you.
1586    
1587    Keep up the good work!
1588    
1589    David
1590    
1591    ----------------------------------------------------------------------
1592    Name: Matthias Blume (blume (at) tti - c (dot) org)
1593    Date: 2004/05/11 14:20:00 CDT
1594    Tag: blume-20040511-installml
1595    Description:
1596    
1597    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
1598    the patch.)
1599    
1600    ----------------------------------------------------------------------
1601    Name: Matthias Blume (blume (at) tti - c (dot) org)
1602    Date: 2004/05/11 14:05:00 CDT
1603    Tag: blume-20040511-nlffi-netbsd
1604    Description:
1605    
1606    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
1607    the patch.)
1608    
1609    ----------------------------------------------------------------------
1610    Name: Matthias Blume (blume (at) tti - c (dot) org)
1611    Date: 2004/05/11 12:05:00 CDT
1612    Tag: blume-20040511-exports
1613    Description:
1614    
1615    As per request by Adam Chlipala <adam@hcoop.net>, extended various
1616    export lists in compiler-related .cm-files.
1617    
1618    ----------------------------------------------------------------------
1619    Name: Matthias Blume (blume (at) tti - c (dot) org)
1620    Date: 2004/05/11 11:35:00 CDT
1621    Tag: blume-20040511-allsource
1622    Description:
1623    
1624    The installer now honors the "src-smlnj" target again, although its meaning
1625    has changed from "all sources required for the compiler" to "all sources
1626    the installer knows about".  In other words, if you enable "src-smlnj"
1627    in the "targets" file, then the installer will pull in sources for
1628    everything.  (Notice that this refers to source code only.  Compiled
1629    code is still only installed for modules that were requested explicitly
1630    or which are required for other modules that were requested explicitly.)
1631    
1632    ----------------------------------------------------------------------
1633    Name: Matthias Blume (blume (at) tti - c (dot) org)
1634    Date: 2004/04/23 17:40:00 CDT
1635    Tag: blume-20040423-ieee-scan
1636    Description:
1637    
1638    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
1639    in the exponent calculation we get INF or ZERO (depending on the mantissa
1640    and the sign of the exponent).
1641    
1642    ----------------------------------------------------------------------
1643    Name: Matthias Blume (blume (at) tti - c (dot) org)
1644    Date: 2004/04/23 10:40:00 CDT
1645    Tag: blume-20040423-ml-build
1646    Description:
1647    
1648    The ml-build script now terminates with a non-0 status when something
1649    goes wrong.
1650    
1651    ----------------------------------------------------------------------
1652    Name: Matthias Blume (blume (at) tti - c (dot) org)
1653    Date: 2004/04/22 16:35:00 CDT
1654    Tag: blume-20040422-Option
1655    Description:
1656    
1657    Made exception Option to be the same as exception Option.Option
1658    (as it should be).
1659    
1660    ----------------------------------------------------------------------
1661    Name: Allen Leung (leunga (at) reservoir (dot) com)
1662    Date: 2004/03/19 14:40:00 EST
1663    Tag: leunga-20040319-cygwin-nlffi
1664    Description:
1665    
1666    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1667    of SML/NJ.  The problem is that
1668    
1669       lib = dlopen(NULL, ...)
1670       f   = dlsym(lib, "malloc");
1671    
1672    does not work on Windows unless we explicitly export symbols
1673    such as 'malloc' during linking.  We fixed this by explicitly
1674    exporting the required symbols with the magic gcc incantation:
1675    
1676        -Wl,--export-all cygwin.def
1677    
1678    where cygwin.def is a file containing all the symbols that
1679    we wish to export.
1680    
1681    I suspect this is a Windows problem and we'll have to
1682    do the same (somehow with windows compilers) when
1683    we build the native win32 version with the system
1684    calls LoadLibrary/GetProcAddress.
1685    
1686    ----------------------------------------------------------------------
1687    Name: Matthias Blume (blume (at) tti - c (dot) org)
1688    Date: 2004/03/04 16:35:00 CST
1689    Tag: blume-20040304-intinf-fmt
1690    Description:
1691    
1692    Fixed problem with IntInf.fmt (sign would show up on the right instead
1693    of on the left for BIN, OCT, and HEX).
1694    
1695    ----------------------------------------------------------------------
1696    Name: Matthias Blume (blume (at) tti - c (dot) org)
1697    Date: 2004/03/04 11:25:00 CST
1698    Tag: blume-20040304-symlinks
1699    Description:
1700    
1701    Fixed problem with installer script (unix only) where bin/ml-yacc and
1702    friends pointed (via symlinks) to absolute locations instead of just
1703    .run-sml.  This was reported by Vesa A Norrman.
1704    
1705    ----------------------------------------------------------------------
1706    Name: Matthias Blume (blume (at) tti - c (dot) org)
1707    Date: 2004/02/13 14:50:00 CST
1708    Tag: Release_110_45
1709    Description:
1710    
1711    New working version (110.45). New bootfiles.
1712    
1713    ----------------------------------------------------------------------
1714    Name: Matthias Blume (blume (at) tti - c (dot) org)
1715    Date: 2004/01/26 15:15:15 CST
1716    Tag: blume-20040126-toplevel
1717    Description:
1718    
1719    Improved handling of exceptions at the interactive toplevel.
1720    
1721    ----------------------------------------------------------------------
1722    Name: Matthias Blume (blume (at) tti - c (dot) org)
1723    Date: 2004/01/26 11:25:00
1724    Tag: blume-20040126-app
1725    Description:
1726    
1727    Type of top-level "app" corrected.
1728    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1729    
1730    ----------------------------------------------------------------------
1731    Name: Matthias Blume (blume (at) tti - c (dot) org)
1732    Date: 2003/11/18 17:10 CST
1733    Tag: blume-20031118-basis-fiddle
1734    Description:
1735    
1736    - changed Timer interface to what might become the spec
1737    - POSIX_FLAGS -> BIT_FLAGS according to spec
1738    - some other minor discrepancies wrt. spec eliminated
1739    
1740    ----------------------------------------------------------------------
1741    Name: Matthias Blume (blume (at) tti - c (dot) org)
1742    Date: 2003/11/06 12:00:00 CST
1743    Tag: Release_110_44
1744    Description:
1745    
1746    New working version (110.44). New bootfiles.
1747    
1748    ----------------------------------------------------------------------
1749    Name: Matthias Blume (blume (at) tti - c (dot) org)
1750    Date: 2003/11/04 11:50:00 CST
1751    Tag: blume-20031104-move-libraries
1752    Description:
1753    
1754    Eliminated the "dont_move_libraries" directive in config/targets.
1755    (The mechanism was broken and could not be fixed easily.  Moreover,
1756    there does not seem to be any reason not to move all libraries into
1757    lib during installation.  I originally implemented this directive as a
1758    backward-compatibility feature when I first introduced the new CM.
1759    Now that things have been stable for a long time and going back to the
1760    old CM is not an option, there is no reason to keep it around.)
1761    
1762    ----------------------------------------------------------------------
1763    Name: Matthias Blume (blume (at) tti - c (dot) org)
1764    Date: 2003/11/03 16:00:00 CST
1765    Tag: blume-20031103-installdir
1766    Description:
1767    
1768    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1769    Richards for pointing out the problem and providing the solution.)
1770    
1771    ----------------------------------------------------------------------
1772    Name: Matthias Blume (blume (at) tti - c (dot) org)
1773    Date: 2003/10/01 17:05:00 CDT
1774    Tag: blume-20031001-lal-mlrisc
1775    Description:
1776    
1777    MLRISC bug fix from Lal.
1778    
1779    ----------------------------------------------------------------------
1780    Name: Matthias Blume (blume (at) tti - c (dot) org)
1781    Date: 2003/09/30 16:10:00 CDT
1782    Tag: blume-20030930-primio-bat
1783    Description:
1784    
1785    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1786    2. Improved .bat files (for Win32 port) to make things work under Win95.
1787       (thanks to Aaron S. Hawley for this one)
1788    
1789    ----------------------------------------------------------------------
1790    Name: Matthias Blume (blume (at) tti - c (dot) org)
1791    Date: 2003/09/26 16:05:00 CDT
1792    Tag: blume-20030926-wrappriv
1793    Description:
1794    
1795    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1796    
1797    ----------------------------------------------------------------------
1798    Name: Matthias Blume (blume (at) tti - c (dot) org)
1799    Date: 2003/09/26 15:00:00 CDT
1800    Tag: blume-20030926-110_43_3
1801    Description:
1802    
1803    - additional cleanup
1804    - version number bump, NEW BOOTFILES
1805    
1806    ----------------------------------------------------------------------
1807    Name: Matthias Blume (blume (at) tti - c (dot) org)
1808    Date: 2003/09/26 12:00:00 CDT
1809    Tag: blume-20030926-ppautoload
1810    Description:
1811    
1812    I modified the read-eval-print loop so that the autoloader gets
1813    invoked whenever the prettyprinter tries to look up a symbol that
1814    is not currently defined in the toplevel environment but which
1815    appears in CM's autoload registry.  As a result, we see far fewer of
1816    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1817    
1818    In addition to this I tried to clean up some pieces of the Basis
1819    implementation (e.g., Socket, Word8Array) in order to prevent other
1820    instances of these ?.Foo.Bar.xxx names from being printed.
1821    
1822    The mechanism that picks names for types still needs some work, though.
1823    (Right now it seems that if there is a type A.t which is defined to
1824    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1825    "?.B.u" although the perhaps more sensible solution would be to use
1826    "A.t" in this case.  In other words, the prettyprinter should follow
1827    a chain of DEFtycs not farther than there are corresponding toplevel
1828    names in the current environment.)
1829    
1830    ----------------------------------------------------------------------
1831    Name: Matthias Blume (blume (at) tti - c (dot) org)
1832    Date: 2003/09/24 16:31:00 CDT
1833    Tag: blume-20030924-installer
1834    Description:
1835    
1836    Another installer tweak:  All the ML code for the installer is now
1837    compiled during CMB.make and put into a little library called
1838    $smlnj/installer.cm.  The installation then simply invokes
1839    
1840       sml -m $smlnj/installer.cm
1841    
1842    and everything happens automagically.
1843    
1844    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1845    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1846            and BINDIR.
1847    
1848    The new scheme guarantees that the ML code responsible for the installation
1849    is in sync with the APIs of the main system.  Also, the installer is
1850    somewhat faster because the installer script is precompiled.
1851    
1852    ----------------------------------------------------------------------
1853    Name: Matthias Blume (blume (at) tti - c (dot) org)
1854    Date: 2003/09/24 15:35:00 CDT
1855    Tag: blume-20030924-synsock
1856    Description:
1857    
1858    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1859    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1860    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1861    implementations of non-blocking operations from CML's Socket
1862    structure.
1863    
1864    ----------------------------------------------------------------------
1865    Name: Matthias Blume (blume (at) tti - c (dot) org)
1866    Date: 2003/09/24 15:10:05 CDT
1867    Tag: blume-20030924-sockets
1868    Description:
1869    
1870    1. Fixed SOCKET API and implementation to match Basis spec.
1871       This required changing the internal representation of sockets to one
1872       that remembers (for each socket file descriptor) whether it is currently
1873       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1874       call is made only if the state actually needs to change).
1875    
1876    2. OS-specific details of sockets were moved into separate files, thus
1877       making it possible to unify the bulk of the socket implementations
1878       between Unix and Win32.
1879    
1880    3. CML's socket API changed accordingly.
1881       (Note that we need to remove non-blocking functions from this API
1882       since they are redundant in the case of CML!)
1883    
1884    4. CML's socket implementation now makes use of non-blocking functions
1885       provided by Basis, thus removing all OS-dependent code from this part
1886       of CML.
1887    
1888    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1889    
1890    ----------------------------------------------------------------------
1891    Name: Matthias Blume (blume (at) tti - c (dot) org)
1892    Date: 2003/09/22 12:10:00 CDT
1893    Tag: blume-20030922-110_43_2
1894    Description:
1895    
1896    Made a new interim version and bootfiles for developer's bootstrapping
1897    convenience.
1898    
1899       110.43.2 -- NEW BOOTFILES
1900    
1901    ----------------------------------------------------------------------
1902    Name: Matthias Blume (blume (at) tti - c (dot) org)
1903    Date: 2003/09/19 15:55:00 CDT
1904    Tag: blume-20030919-cmdir
1905    Description:
1906    
1907    1. new-install.sh -> install.sh
1908    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1909    3. tweaked installer so that another name instead of .cm can be chosen
1910       at install time (by setting the CM_DIR_ARC environment variable
1911       during installation); once installation is complete, the name is
1912       fixed
1913    
1914    ----------------------------------------------------------------------
1915    Name: Matthias Blume (blume (at) tti - c (dot) org)
1916    Date: 2003/09/18 16:00:00 CDT
1917    Tag: blume-20030918-110_43_1
1918    Description:
1919    
1920    Made a new interim version and bootfiles for developer's bootstrapping
1921    convenience.
1922    
1923       110.43.1 -- NEW BOOTFILES
1924    
1925    ----------------------------------------------------------------------
1926    Name: Matthias Blume (blume (at) tti - c (dot) org)
1927    Date: 2003/09/18 15:20:00 CDT
1928    Tag: blume-20030918-misc
1929    Description:
1930    
1931    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1932       this seems to be controversial at the moment) and used those in
1933       Posix.ProcEnv.times.
1934    
1935    2. Added Time.{from,to}Nanoseconds to Time.
1936    
1937    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1938       For example, fromLargeInt never needs to look at more than 3 "big
1939       digits" to get its 53 bits of precision.
1940    
1941    ----------------------------------------------------------------------
1942    Name: Matthias Blume (blume (at) tti - c (dot) org)
1943    Date: 2003/09/17 16:30:00 CDT
1944    Tag: blume-20030917-real32-slices
1945    Description:
1946    
1947    Added an entry to the primitive environment
1948    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1949    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1950    
1951    Removed some of the "magic" constants in real64.sml and replaced them
1952    with code that generates these values from their corresponding
1953    integer counterparts.
1954    
1955    Made all(?) the slice-related changes to the Basis and made everything
1956    compile again...
1957    
1958    ----------------------------------------------------------------------
1959    Name: Matthias Blume (blume (at) tti - c (dot) org)
1960    Date: 2003/09/15 17:45:00 CDT
1961    Tag: blume-20030915-rbase
1962    Description:
1963    
1964    Fixed bug in Real.fromLargeInt.
1965    
1966    ----------------------------------------------------------------------
1967    Name: Matthias Blume (blume (at) tti - c (dot) org)
1968    Date: 2003/09/13 18:11:00 CDT
1969    Tag: blume-20030913-libinstall
1970    Description:
1971    
1972    Minor bugfix in config/libinstall (set anchor with path to
1973    standalone tool after installing it, otherwise libraries that
1974    need ml-lex or ml-yacc won't compile the first time the installer
1975    runs).
1976    
1977    ----------------------------------------------------------------------
1978    Name: Matthias Blume (blume (at) tti - c (dot) org)
1979    Date: 2003/09/12 11:45:00 CDT
1980    Tag: blume-20030912-various
1981    Description:
1982    
1983     - fixed bug in Real.toLargeInt
1984     - fixed bug in Posix.ProcEnv.times
1985     - changed inputLine functions to return an option
1986     - minor installer improvements / bugfixes
1987     - changed default @SMLalloc parameter for x86/celeron to 64k
1988    
1989    ----------------------------------------------------------------------
1990    Name: Matthias Blume (blume (at) tti - c (dot) org)
1991    Date: 2003/09/09 22:00:00 CDT
1992    Tag: Release_110_43
1993    Description:
1994    
1995        New working release 110.43.  New bootfiles.
1996    
1997    ----------------------------------------------------------------------
1998    Name: Matthias Blume (blume (at) tti - c (dot) org)
1999    Date: 2003/09/09 19:20:00 CDT
2000    Tag: blume-20030909-installer
2001    Description:
2002    
2003    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
2004    Modified config/install.bat to take advantage of it.  Also modified
2005    config/install.sh (and called it config/new-install.sh) to take advantage
2006    of it on Unix systems.  (The SML code is (supposed to be) platform-
2007    independent.)
2008    
2009    The installer can now install everything under Win32
2010    as well as under *nix as long as it compiles.
2011    
2012    Other changes:
2013    
2014      - made CML compile again under Win32
2015      - made eXene compile under Win32 (by providing a fake structure UnixSock
2016        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
2017      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
2018        same as type int;  under Win32 it isn't
2019      - fixed some slice-related problems in the win32-specific parts of CML
2020      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
2021        (under Win32, the volume name is case-insensitive, and the
2022        OS.Path code compares volume names for equality)
2023    
2024    ----------------------------------------------------------------------
2025    Name: Matthias Blume (blume (at) tti - c (dot) org)
2026    Date: 2003/09/08 11:55:00 CDT
2027    Tag: blume-20030908-fullpath
2028    Description:
2029    
2030    Made Win32 version of OS.FileSys.fullPath return current directory
2031    when given an empty string.  This is what the spec says, and incidentally,
2032    CM depends on it.  (CM otherwise goes into an infinite loop in certain
2033    cases when presented with the name of a non-existing .cm file.)
2034    
2035    ----------------------------------------------------------------------
2036    Name: Matthias Blume (blume (at) tti - c (dot) org)
2037    Date: 2003/09/04 16:30:00 CDT
2038    Tag: blume-20030905-slices-etc
2039    Description:
2040    
2041      1. Changed interface to vectors and arrays in Basis to match
2042         (draft) Basis spec.
2043      2. Added signatures and implementations of slices according to
2044         Basis spec.
2045      3. Edited source code throughout the system to make it compile again
2046         under 1. and 2.  (In some cases code had to be added to have it
2047         match the new signatures.)
2048      4. MLRISC should be backward-compatible: the copies of the originals
2049         of files that needed to change under 3. were retained, the .cm files
2050         check the compiler version number and use old versions when
2051         appropriate.
2052      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
2053         dirstream -> string option (in accordance with Basis spec).
2054      6. When generating code that counts lines, ml-lex used function
2055         CharVector.foldli, taking advantage of its old interface.
2056         This has been replaced with the corresponding code from
2057         CharVectorSlice.  (html-lex must be re-lexed!)
2058      7. BitArray in smlnj-lib/Util has been extended/modified to match the
2059         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
2060      8. Removed temporary additions (fromInternal, toInternal) from the
2061         (now obsolete) IntInf in smlnj-lib/Util.
2062      9. Cleaned up structure Byte.
2063     10. Added localOffset, scan, and fromString to Date (according to spec).
2064         Cleaned/corrected implementation of Date.
2065         (Still need to check for correctness; implement better canonicalizeDate.)
2066     11. Added "scan" to signature IEEE_REAL.
2067     12. Some improvements to IntInf [in particular: efficiency-hack for
2068         mod and rem when second operand is 2 (for parity checks).]
2069     13. Changed representation of type Time.time, using a single IntInf.int
2070         value counting microseconds.  This considerably simplified the
2071         implementation of structure Time.  We now support negative time
2072         values; scan and fromString handle signs.
2073     14. Functor PrimIO now takes two additional arguments (VectorSlice and
2074         ArraySlice).
2075    
2076    ----------------------------------------------------------------------
2077    Name: Matthias Blume (blume (at) tti - c (dot) org)
2078    Date: 2003/08/28 17:00:00 CDT
2079    Tag: blume-20030828-intinf
2080    Description:
2081    
2082    This is a major update which comes with a version number bump
2083    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
2084    and an implementation of IntInf in the Basis.
2085    
2086    There are a fairly large number of related changes and updates throughout
2087    the system:
2088    
2089    Basis:
2090            - Implemented IntInf.
2091            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
2092            - Added some missing Real64 operations, most notably Real.toLargeInt.
2093            - Added FixedInt as a synonym for Int32.
2094    
2095    compiler:
2096            * Added support for a built-in intinf type.
2097              - literals
2098              - pattern matching
2099              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
2100              - overloading on literals and operations
2101    
2102            This required adding a primitive type intinf, some additional
2103            primops, and implementations for several non-trivial intinf
2104            operations in Core.  (The intinf type is completely abstract
2105            to the compiler; all operations get delegated back to the Core.)
2106    
2107            * Intinf equality is handled by polyequal.  However, the compiler
2108            does not print its usual warning in this case (since polyequal
2109            is the right thing to do there).
2110    
2111            * Improved the organization of structure InlineT.
2112    
2113            * A word about conversion primops:
2114            If conversions involving intinf do not cancel out during
2115            CPS contract, then the compiler must insert calls to Core functions.
2116            Since all core access must be resolved already during the FLINT
2117            translate phase, it would be too late a the time of CPS contract
2118            to add new Core calls.  For this reason, conversion primops
2119            for intinf carry two arguments:  1. the numeric argument that
2120            they are supposed to convert, and 2. the Core function that
2121            can help with this conversion if necessary.  If CPS contract
2122            eliminates a primop, then the associated Core function becomes
2123            dead and goes away.  Intinf conversion primops that do not get
2124            eliminated by CPS contract get rewritten into calls of their
2125            core functions by a separate, new phase.
2126    
2127    interactive system:
2128            - Control.Print.intinfDepth controls max length of intinf constants
2129            being printed. (Analogous to Control.Print.stringDepth.)
2130            - Cleanup in printutil and pputil: got rid of unused stuff and
2131            duplicates; replaced some of the code with code that makes better
2132            use of library functionality.
2133    
2134    CM:
2135            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
2136            are no longer silent.
2137    
2138    CKIT:
2139            Fixed mismatched uses of Int32 and LargeInt.  I always decided
2140            in favor of LargeInt -- which is now the same as IntInf.
2141            CKIT-knowledgable people should check whether this is what's
2142            intended and otherwise change things back to using Int32 or
2143            FixedInt.
2144    
2145    Throughout the code:
2146            Started using IntInf.int literals and built-in operations
2147            (e.g., comparison with 0) where this seems appropriate.
2148    
2149    
2150    ----------------------------------------------------------------------
2151    Name: Dave MacQueen (dbm@cs.uchicago.edu)
2152    Date: 2003/08/13 11:36:00 CDT
2153    Tag: dbm-20030813-mcz-merge1
2154    Description:
2155    
2156    Merging changes from the mcz-branch development branch into trunk.
2157    These changes involve replacement of the emulated old prettyprinter
2158    interface with direct use of the SML/NJ Lib PP library, and fixing
2159    of a couple of bugs (895, 1186) relating to error messages.  A new
2160    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
2161    has been added.
2162    
2163    ----------------------------------------------------------------------
2164    Name: Matthias Blume (blume (at) tti - c (dot) org)
2165    Date: 2003/08/11 15:45:00 CDT
2166    Tag: blume-20030811-windows
2167    Description:
2168    
2169    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
2170    
2171        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
2172    
2173    This patch restores SML/NJ's ability to run under win32.  There are a
2174    number of changes, including fixes for several bugs that had gone
2175    unnoticed until now:
2176    
2177      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
2178        trying to run the win32 version from within cygwin.)
2179    
2180      - There are a number of simple .bat scripts that substitute for their
2181        corresponding Unix shell-scripts.  (See below.)
2182    
2183      - The internals of ml-build have been modified slightly.  The main
2184        difference is that instead of calling ".link-sml" (or link-sml.bat)
2185        using OS.Process.system, the ML process delegates this task back
2186        to the script.  Otherwise problems arise in mixed environments such
2187        as Cygwin where scripts look and work like Unix scripts, but
2188        where OS.Process.system cannot run them.
2189    
2190      - In CM, the srcpath pickler used native pathname syntax -- which
2191        is incorrect in the case of cross-compilation.  The new pickle format
2192        is independent of platform-specific naming conventions.
2193    
2194      - Path configuration files (such as lib/pathconfig) can now choose
2195        between native and standard syntax.  Placing a line of the form
2196    
2197            standard!
2198    
2199        into the file causes all subsequent paths to be interpreted using
2200        CM standard pathname syntax (= Unix conventions); a line
2201    
2202            native!
2203    
2204        switches back to native style.  This was needed so that
2205        path config files can be written portably, see src/system/pathconfig.
2206    
2207      - Runtime system:
2208    
2209          - win32-filesys.c:  get_file_time and set_file_time now
2210            access modification time, not creation time.
2211    
2212          - I/O code made aware of new array representation.
2213    
2214          - Bug fixes in X86.prim.masm.
2215    
2216       - src/system/makeml made aware of win32. (For use under cygwin
2217            and other Unix-environments for windows.)
2218    
2219       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
2220         which caused BinIO.inputAll to fail consistently.
2221    
2222    .bat scripts:
2223    
2224       Windows .bat scripts assume that SMLNJ_HOME is defined.
2225    
2226       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
2227         applications (sml, ml-yacc, ml-lex).
2228       - ml-build.bat: analogous to ml-build.
2229       - config\install.bat:  Analogous to config/install.sh.  This requires
2230         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
2231         (nmake etc. must be on the path, and vcvars32 must have been run.)
2232         Moreover, sources for ml-lex and ml-yacc need to exist under src,
2233         and the bootfile hierarchy must have been unpacked under
2234         sml.boot.x86-win32.
2235         The script is very primitive and does a poor job at error checking.
2236         It only installs the base system, ml-lex, and ml-yacc.  No other
2237         libraries are being installed (i.e., you get only those that
2238         are part of the compiler.)
2239       - link-sml.bat:  analogous to .link-sml, but not currently used
2240    
2241    Unrelated bug fixes:
2242    
2243      - ml-nlffigen now exports structures ST_* corresponding to incomplete
2244        types.
2245      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
2246        otherwise.)
2247    
2248    ----------------------------------------------------------------------
2249    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
2250    Date: 2003/06/17
2251    Tag: macqueen-20030617-bug895
2252    Description:
2253    
2254    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
2255    Tag will be used for new development branch (mcz-branch) for
2256    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
2257    
2258    ----------------------------------------------------------------------
2259    Name: Matthias Blume (blume (at) tti - c (dot) org)
2260    Date: 2003/05/27 16:55:00 CDT
2261    Tag: blume-20030527-polyeq
2262    Description:
2263    
2264    Tried to eliminated most cases of polymorphic equality.
2265    
2266    ----------------------------------------------------------------------
2267    Name: Matthias Blume (blume (at) tti - c (dot) org)
2268    Date: 2003/05/21 17:45:00 CDT
2269    Tag: blume-20030517-complete
2270    Description:
2271    
2272    Two changes:
2273    
2274      1. Added a flag for controlling whether non-exhaustive bindings will
2275         be treated as errors (default is false).
2276      2. Cleaned up the *entire* source tree so that CMB.make goes through
2277         without a single non-exhaustive match- or bind warning.
2278    
2279    ----------------------------------------------------------------------
2280    Name: Matthias Blume (blume (at) tti - c (dot) org)
2281    Date: 2003/05/17 10:20:00 CDT
2282    Tag: blume-20030517-absyn
2283    Description:
2284    
2285    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
2286    
2287       This mainly affects the quality of error messages.  However, some
2288       of the code is now more straightforward than before. (Treatment of
2289       the above four constructs in translate.sml is much simpler than
2290       the "macro-expansion" that was going on before.  Plus, the mach-
2291       compiler no longer gets invoked just to be able to compile an
2292       if-expression.)
2293    
2294    2. The ErrorMsg.Error exception is now caught and absorbed by the
2295       interactive loop.
2296    
2297    ----------------------------------------------------------------------
2298    Name: Allen Leung
2299    Date: 2003/05/16 13:05:00 CDT
2300    Tag: leunga-20030516-cygwin-runtime
2301    Description:
2302    
2303        Ported the runtime system to cygwin, which uses the unix
2304    x86-unix bin files.  Missing/buggy features:
2305    
2306       o getnetbyname, getnetbyaddr: these functions seem to be missing in
2307         the Cygwin library.
2308       o Ctrl-C handling may be flaky.
2309       o Windows system calls and Windows I/O are not supported.
2310    
2311       A new set of binfiles is located at:
2312    
2313         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
2314    
2315       This is only needed for bootstrapping the cygwin version of smlnj.
2316       Other x86 versions can use the existing binfiles.
2317    
2318    ----------------------------------------------------------------------
2319    Name: Matthias Blume
2320    Date: 2003/04/08 15:42:00 CDT
2321    Tag: blume-20030408-listpair
2322    Description:
2323    
2324    1. Added a target 'mlrisc' to installer.
2325    
2326    2. Added missing elements to structure ListPair.
2327    
2328    ----------------------------------------------------------------------
2329    Name: Allen Leung
2330    Date: 2003/01/07 10:40:00 EST
2331    Tag: leunga-20030107-int-rem
2332    Description:
2333    
2334    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
2335    The arguments to the SUBL instruction were swapped.
2336    
2337    ----------------------------------------------------------------------
2338    Name: Matthias Blume
2339    Date: 2002/12/12 16:25:00 EST
2340    Tag: blume-20021212-risc-ra
2341    Description:
2342    
2343    Fixed a serious bug in the rewrite code for FP spilling/reloading that
2344    sent the RA into an infinite loop when floating point registers get
2345    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
2346    110.37 and 110.38.)
2347    There was another set of potential problems related to the handling of
2348    MLRISC annotations (but those did not yet cause real problems, apparently).
2349    
2350    ----------------------------------------------------------------------
2351    Name: Matthias Blume
2352    Date: 2002/12/06 22:40:00 EST
2353    Tag: blume-20021206-cm-fileid
2354    Description:
2355    
2356    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
2357    This fixes the problem of CM getting confused by files that suddenly
2358    change their identity (e.g., by getting unlinked and recreated by some
2359    text editor such as vi).  There might be a better/cheaper/cleaner way
2360    of doing this, but for now this will have to do.
2361    
2362    ----------------------------------------------------------------------
2363    Name: Matthias Blume
2364    Date: 2002/10/28 09:50:00 EST
2365    Tag: blume-20021028-typecheck
2366    Description:
2367    
2368    Exported structure Typecheck from $smlnj/viscomp/core.cm.
2369    
2370    ----------------------------------------------------------------------
2371    Name: Matthias Blume
2372    Date: 2002/10/17 09:10:00 EDT
2373    Tag: Release_110_42
2374    Description:
2375    
2376    In good old tradition, there has been a slight hiccup so that we have
2377    to patch 110.42 after the fact.   The old release tag has been replaced
2378    (see below).
2379    
2380    The change solves a problem with two competing approaches the
2381    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
2382    each other's way.
2383    
2384    This change only affects the runtime system code and the installer script.
2385    (No new bootfiles.)
2386    
2387    ----------------------------------------------------------------------
2388    Name: Matthias Blume
2389    Date: 2002/10/16 12:00:00 EDT
2390    Tag: Release_110_42_removed
2391    Description:
2392    
2393    New working release.  New bootfiles.
2394    
2395    ----------------------------------------------------------------------
2396    Name: Matthias Blume
2397    Date: 2002/10/10 13:10:00 EDT
2398    Tag: blume-20021010-ppc-divs
2399    Description:
2400    
2401    The mltree operator DIVS must be implemented with an overflow check on
2402    the PPC because the hardware indicates divide-by-zero using "overflow" as
2403    well.
2404    
2405    ----------------------------------------------------------------------
2406    Name: Matthias Blume
2407    Date: 2002/07/23 11:45:00
2408    Tag: blume-20020723-smlnj-home
2409    Description:
2410    
2411    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
2412    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
2413    CM_PATHCONFIG is also set), the path configuration file is assumed
2414    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
2415    the entire tree to some other place and everything will "just work".
2416    
2417    (Companion commands such as ml-build and ml-makedepend also sense this
2418    variable.)
2419    
2420    ----------------------------------------------------------------------
2421    Name: Matthias Blume
2422    Date: 2002/07/12 21:19:00 EDT
2423    Tag: blume-20020712-liveness
2424    Description:
2425    
2426    Exported two useful "step" functions from liveness module (MLRISC).
2427    
2428    ----------------------------------------------------------------------
2429    Name: Matthias Blume
2430    Date: 2002/07/05 16:00 EDT
2431    Tag: Release_110_41
2432    Description:
2433    
2434    New working release.  New bootfiles.
2435    
2436    ----------------------------------------------------------------------
2437    Name: Matthias Blume
2438    Date: 2002/07/05 10:25:00 EDT
2439    Tag: blume-20020705-btimp
2440    Description:
2441    
2442    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
2443    other clients can set up backtracing support.
2444    
2445    ----------------------------------------------------------------------
2446    Name: Matthias Blume
2447    Date: 2002/06/25 17:23:00 EDT
2448    Tag: blume-20020625-fpmax
2449    Description:
2450    
2451    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
2452    case.  (The sense of the isNaN test was reversed -- which made min and
2453    max always return their first argument.)
2454    
2455    ----------------------------------------------------------------------
2456    Name: Matthias Blume
2457    Date: 2002/06/11
2458    Tag: blume-20020611-unixpath
2459    Description:
2460    
2461    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
2462    
2463    ----------------------------------------------------------------------
2464    Name: Matthias Blume
2465    Date: 2002/06/10 16:35:00 EDT
2466    Tag: blume-20020610-ieeereal
2467    Description:
2468    
2469    I back-ported my implementation of IEEEReal.fromString from the
2470    idlbasis-devel branch so that we can test it.
2471    
2472    Another small change is that ppDec tries to give more information
2473    than just "<sig>" in the case of functors.  However, this code is
2474    broken in some mysterious way if the functor's body's signature
2475    has not been declared by ascription but gets inferred from the
2476    implementation.  This needs fixing...
2477    
2478    ----------------------------------------------------------------------
2479    Name: Matthias Blume
2480    Date: 2002/05/31
2481    Tag: blume-20020531-btrace-mode
2482    Description:
2483    
2484    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
2485    the wayside when I switched over to using Controls everywhere.)
2486    
2487    ----------------------------------------------------------------------
2488    Name: Lal George
2489    Date: 2002/05/23 12:21:40 EDT
2490    Tag: george-20020523-visual-labels
2491    Description:
2492    
2493    Labels are now displayed in the graphical output to make
2494    the fall-through and target blocks obvious.
2495    
2496    ----------------------------------------------------------------------
2497    Name: Matthias Blume
2498    Date: 2002/05/22 11:03:00 EDT
2499    Tag: blume-20020522-shrink
2500    Description:
2501    
2502    John tweaked yesterday's fix for 1131 to handle an out-of-memory
2503    situation that comes up when allocating huge arrays.
2504    
2505    ----------------------------------------------------------------------
2506    Name: Matthias Blume
2507    Date: 2002/05/21 16:00:00 EDT
2508    Tag: Release_110_40
2509    Description:
2510    
2511    New working release (110.40).  New bootfiles.
2512    
2513    [Also: John Reppy fixed GC bug 1131.]
2514    
2515    ----------------------------------------------------------------------
2516    Name: Matthias Blume
2517    Date: 2002/05/21 12:35:00 EDT
2518    Tag: blume-20020521-cmdoc
2519    Description:
2520    
2521    CM documentation update.
2522    
2523    ----------------------------------------------------------------------
2524    Name: Matthias Blume
2525    Date: 2002/05/21 10:55:00 EDT
2526    Tag: blume-20020521-misc
2527    Description:
2528    
2529    - John tweaked runtime to be silent on heap export (except when
2530      GC messages are on).
2531    - I added a few more things (cross-compiling versions of CMB) to
2532      config/preloads (as suggestions).
2533    
2534    ----------------------------------------------------------------------
2535    Name: Matthias Blume
2536    Date: 2002/05/20 22:25:00 EDT
2537    Tag: blume-20020520-controls
2538    Description:
2539    
2540    - Added ControlUtil structure to control-lib.cm.
2541    - Use it throughout.
2542    - Used Controls facility to define MLRISC controls (as opposed to
2543      registering MLRISC control ref cells with Controls after the fact)
2544    - Fixed messed-up controls priorities.
2545    
2546    * Removed again all the stuff from config/preloads that one wouldn't
2547      be able to preload at the time the initial heap image is built.
2548      (Many libraries, e.g., CML, do not exist yet at this time.  The
2549       only libraries that can be preloaded via config/preloads are those
2550       that come bundled with the bootfiles.)
2551    
2552    ----------------------------------------------------------------------
2553    Name: Matthias Blume
2554    Date: 2002/05/20 10:59:00 EDT
2555    Tag: blume-20020520-preloads
2556    Description:
2557    
2558    Added a lot of commented-out suggestions for things to be included
2559    in config/preloads.
2560    
2561    ----------------------------------------------------------------------
2562    Name: Allen Leung
2563    Date: 2002/05/18 14:20:00 EDT
2564    Tag: leunga-20020518-mdl
2565    Description:
2566    
2567    o Made the mdl tool stuff compile and run again.
2568    o I've disabled all the stuff that depends on RTL specifications; they
2569      are all badly broken anyway.
2570    
2571    ----------------------------------------------------------------------
2572    Name: Matthias Blume
2573    Date: 2002/05/17 16:49:00 EDT
2574    Tag: blume-20020517-controls
2575    Description:
2576    
2577    0. John Reppy made several modifications to the SML/NJ library.
2578       In particular, there is a shiny new controls-lib.cm.
2579    
2580    1. Pushed new controls interface through compiler so that everything
2581       compiles again.
2582    
2583    2. Added FormatComb and FORMAT_COMB to the CML version of the
2584       SML/NJ library (so that CML compiles again).
2585    
2586    3. Modified init scripts because XXX_DEFAULT environment variables
2587       are no longer with us.  (Boot-time initialization is now done
2588       using the same environment variables that are also used for
2589       startup-time initialization of controls.)
2590    
2591    ----------------------------------------------------------------------
2592    Name: Lal George
2593    Date: 2002/05/15 09:20:10 EDT
2594    Tag: george-20020515-pseudo-op-decls
2595    Description:
2596    
2597            All pseudo-ops emitted before the first segment declaration
2598    such as TEXT, DATA, and BSS directives are assumed to be global
2599    declarations and are emitted first in the assembly file. This is
2600    useful in a number of situations where one has pseudo-ops that are not
2601    specific to any segment, and also works around the constraint that one
2602    cannot have client pseudo-ops in the TEXT segment.
2603    
2604            Because no segment is associated with these declarations it is
2605    an error to allocate any space or objects before the first segment
2606    directive and an exception will be raised. However, we cannot make
2607    this check for client pseudo-ops.
2608    
2609            These top level declarations are a field in the CFG graph_info.
2610    In theory you can continue to add to this field after the CFG has been
2611    built -- provided you know what you are doing;-)
2612    
2613    ----------------------------------------------------------------------
2614    Name: Matthias Blume
2615    Date: 2002/05/13 16:40:00 EDT
2616    Tag: blume-20020513-pp-etc
2617    Description:
2618    
2619    A few minor bugfixes:
2620    
2621      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
2622        (Bogus printouts for redefined bindings still occur.  Compiler
2623        bug should no longer occur now.  We need to redo the prettyprinter
2624        from scratch.)
2625    
2626      - CM pathname printer now also adds escape sequences for ( and )
2627    
2628      - commend and docu fixes for ml-nlffi
2629    
2630    ----------------------------------------------------------------------
2631    Name: Matthias Blume
2632    Date: 2002/05/10 16:40:00 EDT
2633    Tag: blume-20020510-erg-textio
2634    Description:
2635    
2636    Applied the following bugfix provided by Emden Gansner:
2637    
2638        Output is corrupted when outputSubstr is used rather than output.
2639    
2640        The problem occurs when a substring
2641    
2642            ss = (s, dataStart, dataLen)
2643    
2644        where dataStart > 0, fills a stream buffer with avail bytes left.
2645        avail bytes of s, starting at index dataStart, are copied into the
2646        buffer, the buffer is flushed, and then the remaining dataLen-avail
2647        bytes of ss are copied into the beginning of the buffer. Instead of
2648        starting this copy at index dataStart+avail in s, the current code
2649        starts the copy at index avail.
2650    
2651        Fix:
2652        In text-io-fn.sml, change line 695 from
2653             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2654        to
2655             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2656    
2657    ----------------------------------------------------------------------
2658    Name: Matthias Blume
2659    Date: 2002/04/12 13:55:00 EDT
2660    Tag: blume-20020412-assyntax
2661    Description:
2662    
2663    1. Grabbed newer assyntax.h from the XFree86 project.
2664    2. Fiddled with how to compile X86.prim.asm without warnings.
2665    3. (Very) Minor cleanup in CM.
2666    
2667    ----------------------------------------------------------------------
2668    Name: Matthias Blume
2669    Date: 2002/04/01 (no joke!) 17:07:00 EST
2670    Tag: blume-20020401-x86div
2671    Description:
2672    
2673    Added full support for div/mod/rem/quot on the x86, using the machine
2674    instruction's two results (without clumsily recomputing the remainder)
2675    directly where appropriate.
2676    
2677    Some more extensive power-of-two support was added to the x86 instruction
2678    selector (avoiding expensive divs, mods, and muls where they can be
2679    replaced with cheaper shifts and masks).  However, this sort of thing
2680    ought to be done earlier, e.g., within the CPS optimizer so that
2681    all architectures benefit from it.
2682    
2683    The compiler compiles to a fixed point, but changes might be somewhat
2684    fragile nevertheless.  Please, report any strange things that you might
2685    see wrt. div/mod/quot/rem...
2686    
2687    ----------------------------------------------------------------------
2688    Name: Matthias Blume
2689    Date: 2002/03/29 17:22:00
2690    Tag: blume-20020329-div
2691    Description:
2692    
2693    Fixed my broken div/mod logic.  Unfortunately, this means that the
2694    inline code for div/mod now has one more comparison than before.
2695    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2696    The problem was with quotient = 0, because that alone does not tell
2697    us which way the rounding went.  One then has to look at whether
2698    remainder and divisor have the same sign...  :(
2699    
2700    Anyway, I replaced the bootfiles with fresh ones...
2701    
2702    ----------------------------------------------------------------------
2703    Name: Matthias Blume
2704    Date: 2002/03/29 14:10:00 EST
2705    Tag: blume-20020329-inlprims
2706    Description:
2707    
2708    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2709    
2710    Primops have changed. This means that the bin/boot-file formats have
2711    changed as well.
2712    
2713    To make sure that there is no confusion, I made a new version.
2714    
2715    
2716    CHANGES:
2717    
2718    * removed REMT from mltree (remainder should never overflow).
2719    
2720    * added primops to deal with divisions of all flavors to the frontend
2721    
2722    * handled these primops all the way through so they map to their respective
2723      MLRISC support
2724    
2725    * used these primops in the implementation of Int, Int32, Word, Word32
2726    
2727    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2728    
2729    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2730    
2731    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2732      floating point
2733    
2734    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2735      and Real (but Real.abs maps to a separate floating-point-only primop)
2736    
2737    
2738    TODO items:
2739    
2740    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2741      instructions because the machine instruction encoder cannot handle
2742      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2743      REM in the first place, and REM is supported by the x86 machine coder.)
2744    
2745    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2746      selection phase.  (The two can be streamlined because the hardware
2747      delivers both quotient and remainder at the same time anyway.)
2748    
2749    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2750      (Currently the behavior is inconsistent both across architectures and
2751      wrt. the draft Basis spec.)
2752    
2753    * Word8 should eventually be handled natively, too.
2754    
2755    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2756      as if there currently is no execution path that could trigger it in
2757      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2758      hold for things like multiplication and division.)
2759    
2760    ----------------------------------------------------------------------
2761    Name: Matthias Blume
2762    Date: 2002/03/27 16:27:00 EST
2763    Tag: blume-20020327-mlrisc-divisions
2764    Description:
2765    
2766    Added support for all four division operations (ML's div, mod, quot,
2767    and rem) to MLRISC.  In the course of doing so, I also rationalized
2768    the naming (no more annoying switch-around of DIV and QUOT), by
2769    parameterizing the operation by div_rounding_mode (which can be either
2770    DIV_TO_ZERO or DIV_TO_NEGINF).
2771    
2772    The generic MLTreeGen functor takes care of compiling all four
2773    operations down to only round-to-zero div.
2774    
2775    Missing pieces:
2776    
2777      * Doing something smarter than relying on MLTreeGen on architectures
2778        like, e.g., the x86 where hardware division delivers both quotient and
2779        remainder at the same time.  With this, the implementation of the
2780        round-to-neginf operations could be further streamlined.
2781    
2782      * Remove inlining support for div/mod/rem from the frontend and replace it
2783        with primops that get carried through to the backend.  Do this for all
2784        int and word types.
2785    
2786    ----------------------------------------------------------------------
2787    Name: Matthias Blume
2788    Date: 2002/03/25 17:25:00 EST
2789    Tag: blume-20020325-divmod
2790    Description:
2791    
2792    I improved (hopefully without breaking them) the implementation of Int.div,
2793    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2794    advantage of the following observations:
2795    
2796      Let  q = x quot y      r = x rem y
2797           d = x div  y      m = x mod y
2798    
2799    where "quot" is the round-to-zero version of integer division that
2800    hardware usually provides.  Then we have:
2801    
2802         r = x - q * y        where neither the * nor the - will overflow
2803         d = if q >= 0 orelse x = q * y then q else q - 1
2804                              where neither the * nor the - will overflow
2805         m = if q >= 0 orelse r = 0 then r else r + y
2806                              where the + will not overflow
2807    
2808    This results in substantial simplification of the generated code.
2809    The following table shows the number of CFG nodes and edges generated
2810    for
2811            fun f (x, y) = x OPER y
2812            (* with OPER \in div, mod, quot, rem *)
2813    
2814    
2815        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2816        --------------------------------------------------------
2817         div |         24 |         39 |         12 |         16
2818         mod |         41 |         71 |         12 |         16
2819        quot |          8 |         10 |          8 |         10
2820         rem |         10 |         14 |          8 |         10
2821    
2822    
2823    ----------------------------------------------------------------------
2824    Name: Matthias Blume
2825    Date: 2002/03/25 22:06:00 EST
2826    Tag: blume-20020325-cprotobug
2827    Description:
2828    
2829    Fixed a bug in cproto (c prototype decoder).
2830    
2831    ----------------------------------------------------------------------
2832    Name: Matthias Blume
2833    Date: 2002/03/25 16:00:00 EST
2834    Tag: blume-20020325-raw-primops
2835    Description:
2836    
2837    I did some cleanup to Allen's new primop code and
2838    replaced yesterday's bootfiles with new ones.
2839    (But they are stored in the same place.)
2840    
2841    ----------------------------------------------------------------------
2842    Name: Matthias Blume
2843    Date: 2002/03/24 22:40:00 EST
2844    Tag: blume-20020324-bootfiles
2845    Description:
2846    
2847    Made the bootfiles that Allen asked for.
2848    
2849    ----------------------------------------------------------------------
2850    Name: Allen Leung
2851    Date: 2002/03/23 15:50:00 EST
2852    Tag: leunga-20020323-flint-cps-rcc-primops
2853    Description:
2854    
2855      1. Changes to FLINT primops:
2856    
2857        (* make a call to a C-function;
2858         * The primop carries C function prototype information and specifies
2859         * which of its (ML-) arguments are floating point. C prototype
2860         * information is for use by the backend, ML information is for
2861         * use by the CPS converter. *)
2862      | RAW_CCALL of { c_proto: CTypes.c_proto,
2863                       ml_args: ccall_type list,
2864                       ml_res_opt: ccall_type option,
2865                       reentrant : bool
2866                     } option
2867       (* Allocate uninitialized storage on the heap.
2868        * The record is meant to hold short-lived C objects, i.e., they
2869        * are not ML pointers.  With the tag, the representation is
2870        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2871        *)
2872      | RAW_RECORD of {tag:bool,sz:int}
2873      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2874    
2875      2.  These CPS primops are now overloaded:
2876    
2877           rawload of {kind:numkind}
2878           rawstore of {kind:numkind}
2879    
2880          The one argument form is:
2881    
2882             rawload {kind} address
2883    
2884          The two argument form is:
2885    
2886             rawload {kind} [ml object, byte-offset]
2887    
2888      3. RAW_CCALL/RCC now takes two extra arguments:
2889    
2890         a. The first is whether the C call is reentrant, i.e., whether
2891            ML state should be saved and restored.
2892         b. The second argument is a string argument specifying the name of
2893            library and the C function.
2894    
2895         These things are currently not handled in the code generator, yet.
2896    
2897      4. In CProto,
2898    
2899         An encoding type of "bool" means "ml object" and is mapped into
2900         C prototype of PTR.  Note that "bool" is different than "string",
2901         even though "string" is also mapped into PTR, because "bool"
2902         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2903    
2904      5. Pickler/unpicker
2905    
2906         Changed to handle RAW_RECORD and newest RAW_CCALL
2907    
2908      6. MLRiscGen,
2909    
2910         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2911         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2912            in the file CodeGen/cpscompile/cps-c-calls.sml
2913    
2914      7. Added the conditional move operator
2915    
2916             condmove of branch
2917    
2918         to cps.  Generation of this is still buggy so it is currently
2919         disabled.
2920    
2921    ----------------------------------------------------------------------
2922    Name: Lal George
2923    Date: 2002/03/22 14:18:25 EST
2924    Tag: george-20020322-cps-branch-prob
2925    Description:
2926    
2927    Implemented the Ball-Larus branch prediction-heuristics, and
2928    incorporated graphical viewers for control flow graphs.
2929    
2930    Ball-Larus Heuristics:
2931    ---------------------
2932    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2933    
2934    By design it uses the Dempster-Shafer theory for combining
2935    probabilities.  For example, in the function:
2936    
2937        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2938    
2939    the ball-larus heuristics predicts that the n=0 is unlikely
2940    (OH-heuristic), and the 'then' branch is unlikely because of the
2941    RH-heuristic -- giving the 'then' branch an even lower combined
2942    probability using the Dempster-Shafer theory.
2943    
2944    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2945    probability of the 'then' branch because of the loop in the else
2946    branch.
2947    
2948    
2949    Graphical Viewing:
2950    ------------------
2951    I merely plugged in Allen's graphical viewers into the compiler. The
2952    additional code is not much. At the top level, saying:
2953    
2954            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2955    
2956    will display the graphical view of the control flow graph just before
2957    back-patching.  daVinci must be in your path for this to work. If
2958    daVinci is not available, then the default viewer can be changed
2959    using:
2960    
2961            Control.MLRISC.getString "viewer"
2962    
2963    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2964    course, these viewers must be in your path.
2965    
2966    The above will display the compilation unit at the level of clusters,
2967    many of which are small, boring, and un-interesting. Also setting:
2968    
2969            Control.MLRISC.getInt "cfg-graphical-view_size"
2970    
2971    will display clusters that are larger than the value set by the above.
2972    
2973    
2974    ----------------------------------------------------------------------
2975    Name: Matthias Blume
2976    Date: 2002/03/21 22:20:00 EST
2977    Tag: blume-20020321-kmp-bugfix
2978    Description:
2979    
2980    Changed the interface to the KMP routine in PreString and fixed
2981    a minor bug in one place where it was used.
2982    
2983    ----------------------------------------------------------------------
2984    Name: Allen Leung
2985    Date: 2002/03/21 20:30:00 EST
2986    Tag: leunga-20020321-cfg
2987    Description:
2988    
2989      Fixed a potential problem in cfg edge splitting.
2990    
2991    ----------------------------------------------------------------------
2992    Name: Allen Leung
2993    Date: 2002/03/21 17:15:00 EST
2994    Tag: leunga-20020321-x86-fp-cfg
2995    Description:
2996    
2997      1. Recoded the buggy parts of x86-fp.
2998    
2999         a. All the block reordering code has been removed.
3000            We now depend on the block placement phases to do this work.
3001    
3002         b. Critical edge splitting code has been simplified and moved into the
3003            CFG modules, as where they belong.
3004    
3005         Both of these were quite buggy and complex.  The code is now much, much
3006         simpler.
3007    
3008      2. X86 backend.
3009    
3010         a. Added instructions for 64-bit support.  Instruction selection for
3011            64-bit has not been committed, however, since that
3012            requires changes to MLTREE which haven't been approved by
3013            Lal and John.
3014    
3015         b. Added support for FUCOMI and FUCOMIP when generating code for
3016            PentiumPro and above.  We only generate these instructions in
3017            the fast-fp mode.
3018    
3019         c. Added cases for JP and JNP in X86FreqProps.
3020    
3021      3. CFG
3022    
3023         CFG now has a bunch of methods for edge splitting and merging.
3024    
3025      4. Machine description.
3026    
3027         John's simplification of MLTREE_BASIS.fcond broke a few machine
3028         description things:
3029    
3030         rtl-build.{sig,sml} and hppa.mdl fixed.
3031    
3032         NOTE: the machine description stuff in the repository is still broken.
3033               Again, I can't put my fixes in because that involves
3034               changes to MLTREE.
3035    
3036    ----------------------------------------------------------------------
3037    Name: Matthias Blume
3038    Date: 2002/03/20 15:55:00 EST
3039    Tag: blume-20020320-kmp
3040    Description:
3041    
3042    Implemented Knuth-Morris-Pratt string matching in PreString and used
3043    it for String.isSubstring, Substring.isSubstring, and
3044    Substring.position.
3045    
3046    (Might need some stress-testing.  Simple examples worked fine.)
3047    
3048    ----------------------------------------------------------------------
3049    Name: Matthias Blume
3050    Date: 2002/03/19 16:37:00 EST
3051    Tag: blume-20020319-witnesses
3052    Description:
3053    
3054    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
3055    
3056    This implements a generic mechanism for changing constness qualifiers
3057    anywhere within big C types without resorting to outright "casts".
3058    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
3059    modify the constness at the outermost level.)
3060    The implementation of "convert" is based on the idea of "witness"
3061    values -- values that are not used by the operation but whose types
3062    "testify" to their applicability.  On the implementation side, "convert"
3063    is simply a projection (returning its second curried argument).  With
3064    cross-module inlining, it should not result in any machine code being
3065    generated.
3066    
3067    ----------------------------------------------------------------------
3068    Name: Matthias Blume
3069    Date: 2002/03/15 16:40:00 EST
3070    Tag: blume-20020315-basis
3071    Description:
3072    
3073    Provided (preliminary?) implementations for
3074    
3075      {String,Substring}.{concatWith,isSuffix,isSubstring}
3076    
3077    and
3078    
3079      Substring.full
3080    
3081    Those are in the Basis spec but they were missing in SML/NJ.
3082    
3083    ----------------------------------------------------------------------
3084    Name: Matthias Blume
3085    Date: 2002/03/14 21:30:00 EST
3086    Tag: blume-20020314-controls
3087    Description:
3088    
3089    Controls:
3090    ---------
3091    
3092    1. Factored out the recently-added Controls : CONTROLS stuff and put
3093       it into its own library $/controls-lib.cm.  The source tree for
3094       this is under src/smlnj-lib/Controls.
3095    
3096    2. Changed the names of types and functions in this interface, so they
3097       make a bit more "sense":
3098    
3099          module -> registry
3100          'a registry -> 'a group
3101    
3102    3. The interface now deals in ref cells only.  The getter/setter interface
3103       is (mostly) gone.
3104    
3105    4. Added a function that lets one register an already-existing ref cell.
3106    
3107    5. Made the corresponding modifications to the rest of the code so that
3108       everything compiles again.
3109    
3110    6. Changed the implementation of Controls.MLRISC back to something closer
3111       to the original.  In particular, this module (and therefore MLRISC)
3112       does not depend on Controls.  There now is some link-time code in
3113       int-sys.sml that registers the MLRISC controls with the Controls
3114       module.
3115    
3116    CM:
3117    ---
3118    
3119      * One can now specify the lambda-split aggressiveness in init.cmi.
3120    
3121    ----------------------------------------------------------------------
3122    Name: Allen Leung
3123    Date: 2002/03/13 17:30:00 EST
3124    Tag: leunga-20020313-x86-fp-unary
3125    Description:
3126    
3127    Bug fix for:
3128    
3129    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
3130    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
3131    > - fun f(x,(y,z)) = Real.~ y;
3132    > [autoloading]
3133    > [autoloading done]
3134    >       fchsl   (%eax), 184(%esp)
3135    > Error: MLRisc bug: X86MCEmitter.emitInstr
3136    >
3137    > uncaught exception Error
3138    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
3139    
3140    The problem was that the code generator did not generate any fp registers
3141    in this case, and the ra didn't know that it needed to run the X86FP phase to
3142    translate the pseudo fp instruction.   This only happened with unary fp
3143    operators in certain situations.
3144    
3145    ----------------------------------------------------------------------
3146    Name: Matthias Blume
3147    Date: 2002/03/13 14:00:00 EST
3148    Tag: blume-20020313-overload-etc
3149    Description:
3150    
3151    1. Added _overload as a synonym for overload for backward compatibility.
3152       (Control.overloadKW must be true for either version to be accepted.)
3153    
3154    2. Fixed bug in install script that caused more things to be installed
3155       than what was requested in config/targets.
3156    
3157    3. Made CM aware of the (_)overload construct so that autoloading
3158       works.
3159    
3160    ----------------------------------------------------------------------
3161    Name: Matthias Blume
3162    Date: 2002/03/12 22:03:00 EST
3163    Tag: blume-20020312-url
3164    Description:
3165    
3166    Forgot to update BOOT and srcarchiveurl.
3167    
3168    ----------------------------------------------------------------------
3169    Name: Matthias Blume
3170    Date: 2002/03/12 17:30:00 EST
3171    Tag: blume-20020312-version110392
3172    Description:
3173    
3174    Yet another version number bump (because of small changes to the
3175    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
3176    
3177    Changes:
3178    
3179      The new pid generation scheme described a few weeks ago was overly
3180      complicated.  I implemented a new mechanism that is simpler and
3181      provides a bit more "stability":  Once CM has seen a compilation
3182      unit, it keeps its identity constant (as long as you do not delete
3183      those crucial CM/GUID/* files).  This means that when you change
3184      an interface, compile, then go back to the old interface, and
3185      compile again, you arrive at the original pid.
3186    
3187      There now also is a mechanism that instructs CM to use the plain
3188      environment hash as a module's pid (effectively making its GUID
3189      the empty string).  For this, "noguid" must be specified as an
3190      option to the .sml file in question within its .cm file.
3191      This is most useful for code that is being generated by tools such
3192      as ml-nlffigen (because during development programmers tend to
3193      erase the tool's entire output directory tree including CM's cached
3194      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
3195      revert to the old, broken behavior of SML/NJ, but in specific cases
3196      where there is no danger of interface confusion, its use is ok
3197      (I think).
3198    
3199      ml-nlffigen by default generates "noguid" annotations.  They can be
3200      turned off by specifying -guid in its command line.
3201    
3202    ----------------------------------------------------------------------
3203    Name: Lal George
3204    Date: 2002/03/12 12 14:42:36 EST
3205    Tag: george-20020312-frequency-computation
3206    Description:
3207    
3208    Integrated jump chaining and static block frequency into the
3209    compiler. More details and numbers later.
3210    
3211    ----------------------------------------------------------------------
3212    Name: Lal George
3213    Date: 2002/03/11 11 22:38:53 EST
3214    Tag: george-20020311-jump-chain-elim
3215    Description:
3216    
3217    Tested the jump chain elimination on all architectures (except the
3218    hppa).  This is on by default right now and is profitable for the
3219    alpha and x86, however, it may not be profitable for the sparc and ppc
3220    when compiling the compiler.
3221    
3222    The gc test will typically jump to a label at the end of the cluster,
3223    where there is another jump to an external cluster containing the actual
3224    code to invoke gc. This is to allow factoring of common gc invocation
3225    sequences. That is to say, we generate:
3226    
3227            f:
3228               testgc
3229               ja   L1      % jump if above to L1
3230    
3231            L1:
3232               jmp L2
3233    
3234    
3235    After jump chain elimination the 'ja L1' instructions is converted to
3236    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
3237    up being implemented in their long form (if L2 is far away) using:
3238    
3239            jbe     L3      % jump if below or equal to L3
3240            jmp     L2
3241         L3:
3242            ...
3243    
3244    
3245    For large compilation units L2  may be far away.
3246    
3247    
3248    ----------------------------------------------------------------------
3249    Name: Matthias Blume
3250    Date: 2002/03/11 13:30:00 EST
3251    Tag: blume-20020311-mltreeeval
3252    Description:
3253    
3254    A functor parameter was missing.
3255    
3256    ----------------------------------------------------------------------
3257    Name: Allen Leung
3258    Date: 2002/03/11 10:30:00 EST
3259    Tag: leunga-20020311-runtime-string0
3260    Description:
3261    
3262       The representation of the empty string now points to a
3263    legal null terminated C string instead of unit.  It is now possible
3264    to convert an ML string into C string with InlineT.CharVector.getData.
3265    This compiles into one single machine instruction.
3266    
3267    ----------------------------------------------------------------------
3268    Name: Allen Leung
3269    Date: 2002/03/10 23:55:00 EST
3270    Tag: leunga-20020310-x86-call
3271    Description:
3272    
3273       Added machine generation for CALL instruction (relative displacement mode)
3274    
3275    ----------------------------------------------------------------------
3276    Name: Matthias Blume
3277    Date: 2002/03/08 16:05:00
3278    Tag: blume-20020308-entrypoints
3279    Description:
3280    
3281    Version number bumped to 110.39.1.  NEW BOOTFILES!
3282    
3283    Entrypoints: non-zero offset into a code object where execution should begin.
3284    
3285    - Added the notion of an entrypoint to CodeObj.
3286    - Added reading/writing of entrypoint info to Binfile.
3287    - Made runtime system bootloader aware of entrypoints.
3288    - Use the address of the label of the first function given to mlriscGen
3289      as the entrypoint.  This address is currently always 0, but it will
3290      not be 0 once we turn on block placement.
3291    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
3292      with entry points) from mlriscGen.
3293    
3294    ----------------------------------------------------------------------
3295    Name: Allen Leung
3296    Date: 2002/03/07 20:45:00 EST
3297    Tag: leunga-20020307-x86-cmov
3298    Description:
3299    
3300       Bug fixes for CMOVcc on x86.
3301    
3302       1. Added machine code generation for CMOVcc
3303       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
3304       3. CMOVcc cannot have an immediate operand as argument.
3305    
3306    ----------------------------------------------------------------------
3307    Name: Matthias Blume
3308    Date: 2002/03/07 16:15:00 EST
3309    Tag: blume-20020307-controls
3310    Description:
3311    
3312    This is a very large but mostly boring patch which makes (almost)
3313    every tuneable compiler knob (i.e., pretty much everything under
3314    Control.* plus a few other things) configurable via both the command
3315    line and environment variables in the style CM did its configuration
3316    until now.
3317    
3318    Try starting sml with '-h' (or, if you are brave, '-H')
3319    
3320    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
3321    implements the underlying generic mechanism.
3322    
3323    The interface to some of the existing such facilities has changed somewhat.
3324    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
3325    (The getFoo interface is still there for backward-compatibility, but its
3326    use is deprecated.)
3327    
3328    The ml-build script passes -Cxxx=yyy command-line arguments through so
3329    that one can now twiddle the compiler settings when using this "batch"
3330    compiler.
3331    
3332    TODO items:
3333    
3334    We should go through and throw out all controls that are no longer
3335    connected to anything.  Moreover, we should go through and provide
3336    meaningful (and correct!) documentation strings for those controls
3337    that still are connected.
3338    
3339    Currently, multiple calls to Controls.new are accepted (only the first
3340    has any effect).  Eventually we should make sure that every control
3341    is being made (via Controls.new) exactly once.  Future access can then
3342    be done using Controls.acc.
3343    
3344    Finally, it would probably be a good idea to use the getter-setter
3345    interface to controls rather than ref cells.  For the time being, both
3346    styles are provided by the Controls module, but getter-setter pairs are
3347    better if thread-safety is of any concern because they can be wrapped.
3348    
3349    *****************************************
3350    
3351    One bug fix: The function blockPlacement in three of the MLRISC
3352    backpatch files used to be hard-wired to one of two possibilities at
3353    link time (according to the value of the placementFlag).  But (I
3354    think) it should rather sense the flag every time.
3355    
3356    *****************************************
3357    
3358    Other assorted changes (by other people who did not supply a HISTORY entry):
3359    
3360    1. the cross-module inliner now works much better (Monnier)
3361    2. representation of weights, frequencies, and probabilities in MLRISC
3362       changed in preparation of using those for weighted block placement
3363       (Reppy, George)
3364    
3365    ----------------------------------------------------------------------
3366    Name: Lal George
3367    Date: 2002/03/07 14:44:24 EST 2002
3368    Tag: george-20020307-weighted-block-placement
3369    
3370    Tested the weighted block placement optimization on all architectures
3371    (except the hppa) using AMPL to generate the block and edge frequencies.
3372    Changes were required in the machine properties to correctly
3373    categorize trap instructions. There is an MLRISC flag
3374    "weighted-block-placement" that can be used to enable weighted block
3375    placement, but this will be ineffective without block/edge
3376    frequencies (coming soon).
3377    
3378    
3379    ----------------------------------------------------------------------
3380    Name: Lal George
3381    Date: 2002/03/05 17:24:48 EST
3382    Tag: george-20020305-linkage-cluster
3383    
3384    In order to support the block placement optimization, a new cluster
3385    is generated as the very first cluster (called the linkage cluster).
3386    It contains a single jump to the 'real' entry point for the compilation
3387    unit. Block placement has no effect on the linkage cluster itself, but
3388    all the other clusters  have full freedom in the manner in which they
3389    reorder blocks or functions.
3390    
3391    On the x86 the typical linkage code that is generated is:
3392       ----------------------
3393            .align 2
3394       L0:
3395            addl    $L1-L0, 72(%esp)
3396            jmp     L1
3397    
3398    
3399            .align  2
3400       L1:
3401       ----------------------
3402    
3403    72(%esp) is the memory location for the stdlink register. This
3404    must contain the address of the CPS function being called. In the
3405    above example, it contains the address of  L0; before
3406    calling L1 (the real entry point for the compilation unit), it
3407    must contain the address for L1, and hence
3408    
3409            addl $L1-L0, 72(%esp)
3410    
3411    I have tested this on all architectures except the hppa.The increase
3412    in code size is of course negligible
3413    
3414    ----------------------------------------------------------------------
3415    Name: Allen Leung
3416    Date: 2002/03/03 13:20:00 EST
3417    Tag: leunga-20020303-mlrisc-tools
3418    
3419      Added #[ ... ] expressions to mlrisc tools
3420    
3421    ----------------------------------------------------------------------
3422    Name: Matthias Blume
3423    Date: 2002/02/27 12:29:00 EST
3424    Tag: blume-20020227-cdebug
3425    Description:
3426    
3427    - made types in structure C and C_Debug to be equal
3428    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
3429    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
3430    
3431    ----------------------------------------------------------------------
3432    Name: Matthias Blume
3433    Date: 2002/02/26 12:00:00 EST
3434    Tag: blume-20020226-ffi
3435    Description:
3436    
3437    1. Fixed a minor bug in CM's "noweb" tool:
3438       If numbering is turned off, then truly don't number (i.e., do not
3439       supply the -L option to noweb).  The previous behavior was to supply
3440       -L'' -- which caused noweb to use the "default" line numbering scheme.
3441       Thanks to Chris Richards for pointing this out (and supplying the fix).
3442    
3443    2. Once again, I reworked some aspects of the FFI:
3444    
3445       A. The incomplete/complete type business:
3446    
3447       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
3448         gone!
3449       - ML types representing an incomplete type are now *equal* to
3450         ML types representing their corresponding complete types (just like
3451         in C).  This is still safe because ml-nlffigen will not generate
3452         RTTI for incomplete types, nor will it generate functions that
3453         require access to such RTTI.   But when ML code generated from both
3454         incomplete and complete versions of the C type meet, the ML types
3455         are trivially interoperable.
3456    
3457         NOTE:  These changes restore the full generality of the translation
3458         (which was previously lost when I eliminated functorization)!
3459    
3460       B. Enum types:
3461    
3462       - Structure C now has a type constructor "enum" that is similar to
3463         how the "su" constructor works.  However, "enum" is not a phantom
3464         type because each "T enum" has values (and is isomorphic to
3465         MLRep.Signed.int).
3466       - There are generic access operations for enum objects (using
3467         MLRep.Signed.int).
3468       - ml-nlffigen will generate a structure E_foo for each "enum foo".
3469         * The structure contains the definition of type "mlrep" (the ML-side
3470         representation type of the enum).  Normally, mlrep is the same
3471         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
3472         then mlrep will be defined as a datatype -- thus facilitating
3473         pattern matching on mlrep values.
3474         ("-ec" will be suppressed if there are duplicate values in an
3475          enumeration.)
3476         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
3477         will be generated for each C enum constant xxx.
3478         * Conversion functions m2i and i2m convert between mlrep and
3479         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
3480         * Coversion functions c and ml convert between mlrep and "tag enum".
3481         * Access functions (get/set) fetch and store mlrep values.
3482       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
3483         enumerations are merged into one single enumeration represented by
3484         structure E_'.
3485    
3486    ----------------------------------------------------------------------
3487    Name: Allen Leung
3488    Date: 2002/02/25 04:45:00 EST
3489    Tag: leunga-20020225-cps-spill
3490    
3491    This is a new implementation of the CPS spill phase.
3492    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
3493    In case of problems, replace it with the old file spill.sml
3494    
3495    The current compiler runs into some serious performance problems when
3496    constructing a large record.  This can happen when we try to compile a
3497    structure with many items.  Even a very simple structure like the following
3498    makes the compiler slow down.
3499    
3500        structure Foo = struct
3501           val x_1 = 0w1 : Word32.int
3502           val x_2 = 0w2 : Word32.int
3503           val x_3 = 0w3 : Word32.int
3504           ...
3505           val x_N = 0wN : Word32.int
3506        end
3507    
3508    The following table shows the compile time, from N=1000 to N=4000,
3509    with the old compiler:
3510    
3511    N
3512    1000   CPS 100 spill                           0.04u  0.00s  0.00g
3513           MLRISC ra                               0.06u  0.00s  0.05g
3514              (spills = 0 reloads = 0)
3515           TOTAL                                   0.63u  0.07s  0.21g
3516    
3517    1100   CPS 100 spill                           8.25u  0.32s  0.64g
3518           MLRISC ra                               5.68u  0.59s  3.93g
3519              (spills = 0 reloads = 0)
3520           TOTAL                                   14.71u  0.99s  4.81g
3521    
3522    1500   CPS 100 spill                           58.55u  2.34s  1.74g
3523           MLRISC ra                               5.54u  0.65s  3.91g
3524              (spills = 543 reloads = 1082)
3525           TOTAL                                   65.40u  3.13s  6.00g
3526    
3527    2000   CPS 100 spill                           126.69u  4.84s  3.08g
3528           MLRISC ra                               0.80u  0.10s  0.55g
3529              (spills = 42 reloads = 84)
3530           TOTAL                                   129.42u  5.10s  4.13g
3531    
3532    3000   CPS 100 spill                           675.59u  19.03s  11.64g
3533           MLRISC ra                               2.69u  0.27s  1.38g
3534              (spills = 62 reloads = 124)
3535           TOTAL                                   682.48u  19.61s  13.99g
3536    
3537    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
3538           MLRISC ra                               4.96u  0.27s  2.72g
3539              (spills = 85 reloads = 170)
3540           TOTAL                                   2375.26u  57.21s  48.00g
3541    
3542    As you can see the old cps spill module suffers from some serious
3543    performance problem.  But since I cannot decipher the old code fully,
3544    instead of patching the problems up, I'm reimplementing it
3545    with a different algorithm.  The new code is more modular,
3546    smaller when compiled, and substantially faster
3547    (O(n log n) time and O(n) space).  Timing of the new spill module:
3548    
3549    4000  CPS 100 spill                           0.02u  0.00s  0.00g
3550          MLRISC ra                               0.25u  0.02s  0.15g
3551             (spills=1 reloads=3)
3552          TOTAL                                   7.74u  0.34s  1.62g
3553    
3554    Implementation details:
3555    
3556    As far as I can tell, the purpose of the CPS spill module is to make sure the
3557    number of live variables at any program point (the bandwidth)
3558    does not exceed a certain limit, which is determined by the
3559    size of the spill area.
3560    
3561    When the bandwidth is too large, we decrease the register pressure by
3562    packing live variables into spill records.  How we achieve this is
3563    completely different than what we did in the old code.
3564    
3565    First, there is something about the MLRiscGen code generator
3566    that we should be aware of:
3567    
3568    o MLRiscGen performs code motion!
3569    
3570       In particular, it will move floating point computations and
3571       address computations involving only the heap pointer to
3572       their use sites (if there is only a single use).
3573       What this means is that if we have a CPS record construction
3574       statement
3575    
3576           RECORD(k,vl,w,e)
3577    
3578       we should never count the new record address w as live if w
3579       has only one use (which is often the case).
3580    
3581       We should do something similar to floating point, but the transformation
3582       there is much more complex, so I won't deal with that.
3583    
3584    Secondly, there are now two new cps primops at our disposal:
3585    
3586     1. rawrecord of record_kind option
3587        This pure operator allocates some uninitialized storage from the heap.
3588        There are two forms:
3589    
3590         rawrecord NONE [INT n]  allocates a tagless record of length n
3591         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
3592                                     and initializes the tag.
3593    
3594     2. rawupdate of cty
3595          rawupdate cty (v,i,x)
3596          Assigns to x to the ith component of record v.
3597          The storelist is not updated.
3598    
3599    We use these new primops for both spilling and increment record construction.
3600    
3601     1. Spilling.
3602    
3603        This is implemented with a linear scan algorithm (but generalized
3604        to trees).  The algorithm will create a single spill record at the
3605        beginning of the cps function and use rawupdate to spill to it,
3606        and SELECT or SELp to reload from it.  So both spills and reloads
3607        are fine-grain operations.  In contrast, in the old algorithm
3608        "spills" have to be bundled together in records.
3609    
3610        Ideally, we should sink the spill record construction to where
3611        it is needed.  We can even split the spill record into multiple ones
3612        at the places where they are needed.  But CPS is not a good
3613        representation for global code motion, so I'll keep it simple and
3614        am not attempting this.
3615    
3616     2. Incremental record construction (aka record splitting).
3617    
3618        Long records with many component values which are simulatenously live
3619        (recall that single use record addresses are not considered to
3620         be live) are constructed with rawrecord and rawupdate.
3621        We allocate space on the heap with rawrecord first, then gradually
3622        fill it in with rawupdate.  This is the technique suggested to me
3623        by Matthias.
3624    
3625        Some restrictions on when this is applicable:
3626        1. It is not a VECTOR record.  The code generator currently does not handle
3627           this case. VECTOR record uses double indirection like arrays.
3628        2. All the record component values are defined in the same "basic block"
3629           as the record constructor.  This is to prevent speculative
3630           record construction.
3631    
3632    ----------------------------------------------------------------------
3633    Name: Allen Leung
3634    Date: 2002/02/22 01:02:00 EST
3635    Tag: leunga-20020222-mlrisc-tools
3636    
3637    Minor bug fixes in the parser and rewriter
3638    
3639    ----------------------------------------------------------------------
3640    Name: Allen Leung
3641    Date: 2002/02/21 20:20:00 EST
3642    Tag: leunga-20020221-peephole
3643    
3644    Regenerated the peephole files.  Some contained typos in the specification
3645    and some didn't compile because of pretty printing bugs in the old version
3646    of 'nowhere'.
3647    
3648    ----------------------------------------------------------------------
3649    Name: Allen Leung
3650    Date: 2002/02/19 20:20:00 EST
3651    Tag: leunga-20020219-mlrisc-tools
3652    Description:
3653    
3654       Minor bug fixes to the mlrisc-tools library:
3655    
3656       1.  Fixed up parsing colon suffixed keywords
3657       2.  Added the ability to shut the error messages up
3658       3.  Reimplemented the pretty printer and fixed up/improved
3659           the pretty printing of handle and -> types.
3660       4.  Fixed up generation of literal symbols in the nowhere tool.
3661       5.  Added some SML keywords to to sml.sty
3662    
3663    ----------------------------------------------------------------------
3664    Name: Matthias Blume
3665    Date: 2002/02/19 16:20:00 EST
3666    Tag: blume-20020219-cmffi
3667    Description:
3668    
3669    A wild mix of changes, some minor, some major:
3670    
3671    * All C FFI-related libraries are now anchored under $c:
3672        $/c.cm      --> $c/c.cm
3673        $/c-int.cm  --> $c/internals/c-int.cm
3674        $/memory.cm --> $c/memory/memory.cm
3675    
3676    * "make" tool (in CM) now treats its argument pathname slightly
3677      differently:
3678        1. If the native expansion is an absolute name, then before invoking
3679           the "make" command on it, CM will apply OS.Path.mkRelative
3680           (with relativeTo = OS.FileSys.getDir()) to it.
3681        2. The argument will be passed through to subsequent phases of CM
3682           processing without "going native".  In particular, if the argument
3683           was an anchored path, then "make" will not lose track of that anchor.
3684    
3685    * Compiler backends now "know" their respective C calling conventions
3686      instead of having to be told about it by ml-nlffigen.  This relieves
3687      ml-nlffigen from one of its burdens.
3688    
3689    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3690    
3691    * Export C_DEBUG and C_Debug from $c/c.cm.
3692    
3693    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3694      subtyping relationship between incomplete pointers and their complete
3695      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3696      with the convention of instantiating 'o with ('t, 'c) obj whenever
3697      the pointer target type is complete.  In the incomplete case, 'o
3698      will be instantiated with some "'c iobj" -- a type obtained by
3699      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3700    
3701      Operations that work on both incomplete and complete pointer types are
3702      typed as taking an 'o ptr while operations that require the target to
3703      be known are typed as taking some ('t, 'c) obj ptr.
3704    
3705      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3706      where void is an eqtype without any values.  This makes it possible
3707      to work on voidptr values using functions meant to operate on light
3708      incomplete pointers.
3709    
3710    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3711      been vastly simplified.
3712    
3713    ----------------------------------------------------------------------
3714    Name: Matthias Blume
3715    Date: 2002/02/19 10:48:00 EST
3716    Tag: blume-20020219-pqfix
3717    Description:
3718    
3719    Applied Chris Okasaki's bug fix for priority queues.
3720    
3721    ----------------------------------------------------------------------
3722    Name: Matthias Blume
3723    Date: 2002/02/15 17:05:00
3724    Tag: Release_110_39
3725    Description:
3726    
3727    Last-minute retagging is becoming a tradition... :-(
3728    
3729    This is the working release 110.39.
3730    
3731    ----------------------------------------------------------------------
3732    Name: Matthias Blume
3733    Date: 2002/02/15 16:00:00 EST
3734    Tag: Release_110_39-orig
3735    Description:
3736    
3737    Working release 110.39.  New bootfiles.
3738    
3739    (Update: There was a small bug in the installer so it wouldn't work
3740    with all shells.  So I retagged. -Matthias)
3741    
3742    ----------------------------------------------------------------------
3743    Name: Matthias Blume
3744    Date: 2002/02/15 14:17:00 EST
3745    Tag: blume-20020215-showbindings
3746    Description:
3747    
3748    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3749    the latter can be useful for exploring what bindings are available at
3750    the interactive prompt.  (The first function returns only the list
3751    of symbols that are really bound, the second prints those but also the
3752    ones that CM's autoloading mechanism knows about.)
3753    
3754    ----------------------------------------------------------------------
3755    Name: Matthias Blume
3756    Date: 2002/02/15 12:08:00 EST
3757    Tag: blume-20020215-iptrs
3758    Description:
3759    
3760    Two improvements to ml-nlffigen:
3761    
3762      1. Write files only if they do not exist or if their current contents
3763         do not coincide with what's being written.  (That is, avoid messing
3764         with the time stamps unless absolutely necessary.)
3765    
3766      2. Implement a "repository" mechanism for generated files related
3767         to "incomplete pointer types".   See the README file for details.
3768    
3769    ----------------------------------------------------------------------
3770    Name: Matthias Blume
3771    Date: 2002/02/14 11:50:00 EST
3772    Tag: blume-20020214-quote
3773    Description:
3774    
3775    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3776    because of the new and improved tag generation scheme.  (Thanks to Allen
3777    Leung for pointing it out.)
3778    
3779    ----------------------------------------------------------------------
3780    Name: Lal George
3781    Date: 2002/02/14 09:55:27 EST 2002
3782    Tag: george-20020214-isabelle-bug
3783    Description:
3784    
3785    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3786    of Isabelle on the x86.
3787    
3788    From Allen:
3789    -----------
3790     I've found the problem:
3791    
3792         in ra-core.sml, I use the counter "blocked" to keep track of the
3793         true number of elements in the freeze queue.  When the counter goes
3794         to zero, I skip examining the queue.  But I've messed up the
3795         bookkeeping in combine():
3796    
3797             else ();
3798             case !ucol of
3799               PSEUDO => (if !cntv > 0 then
3800                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3801                                        ^^^^^^^^^^^^^^^^^^^^^^^
3802                      moveu := mergeMoveList(!movev, !moveu)
3803                     )
3804                  else ();
3805    
3806         combine() is called to coalesce two nodes u and v.
3807         I think I was thinking that if the move counts of u and v are both
3808         greater than zero then after they are coalesced then one node is
3809         removed from the freeze queue.  Apparently I was thinking that
3810         both u and v are of low degree, but that's clearly not necessarily true.
3811    
3812    
3813    02/12/2002:
3814        Here's the patch.  HOL now compiles.
3815    
3816        I don't know how this impact on performance (compile
3817        time or runtime).  This bug caused the RA (especially on the x86)
3818        to go thru the potential spill phase when there are still nodes on the
3819        freeze queue.
3820    
3821    
3822    
3823    
3824    ----------------------------------------------------------------------
3825    Name: Matthias Blume
3826    Date: 2002/02/13 22:40:00 EST
3827    Tag: blume-20020213-fptr-rtti
3828    Description:
3829    
3830    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3831    updates.
3832    
3833    ----------------------------------------------------------------------
3834    Name: Matthias Blume
3835    Date: 2002/02/13 16:41:00 EST
3836    Tag: blume-20020213-cmlpq
3837    Description:
3838    
3839    Added new priority queue export symbols (which have just been added to
3840    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3841    not compile and the installer would choke.)
3842    
3843    ----------------------------------------------------------------------
3844    Name: Matthias Blume
3845    Date: 2002/02/13 16:15:00 EST
3846    Tag: blume-20020213-various
3847    Description:
3848    
3849    1. More tweaks to ml-nlffigen:
3850    
3851       - better internal datastructures (resulting in slight speedup)
3852       - "-match" option requires exact match
3853       - "localized" gensym counters (untagged structs/unions nested within
3854         other structs/unions or within typedefs get a fresh counter; their
3855         tag will be prefixed by a concatenation of their parents' tags)
3856       - bug fixes (related to calculation of transitive closure of types
3857         to be included in the output)
3858    
3859    2. Minor Basis updates:
3860    
3861       - added implementations for List.collate and Option.app
3862    
3863    ----------------------------------------------------------------------
3864    Name: Matthias Blume
3865    Date: 2002/02/11 15:55:00 EST
3866    Tag: blume-20020211-gensym
3867    Description:
3868    
3869    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3870    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3871    names (ML structure names that correspond to unnamed C structs, unions,
3872    and enums), so that separate runs of ml-nlffigen do not clash.
3873    
3874    ----------------------------------------------------------------------
3875    Name: Matthias Blume
3876    Date: 2002/02/11 12:05:00 EST
3877    Tag: blume-20020211-gensml
3878    Description:
3879    
3880    A quick fix for a problem with GenSML (in the pgraph-util library):
3881    Make generation of toplevel "local" optional.  (Strictly speaking,
3882    signature definitions within "local" are not legal SML.)
3883    
3884    Other than that: updates to INSTALL and cm/TODO.
3885    
3886    ----------------------------------------------------------------------
3887    Name: Matthias Blume
3888    Date: 2002/02/08 15:00:00 EST
3889    Tag: blume-20020208-uniquepid
3890    Description:
3891    
3892    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3893    
3894    1. The installer (config/install.sh) has gotten smarter:
3895    
3896         - Configuration options are a bit easier to specify now
3897           (in config/targets).
3898         - Bug in recognizing .tar.bz2 files fixed.
3899         - Installer automatically resolves dependencies between
3900           configuration options (e.g., if you ask for eXene, you will
3901           also get cml -- regardless whether you asked for it or not).
3902         - Installer can run in "quieter mode" by setting the environment
3903           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3904           "completely silent", though.
3905         - Build HashCons library as part of smlnj-lib.
3906    
3907    2. A new scheme for assigning persistent identifiers to compilation
3908       units (and, by extension, to types etc.) has been put into place.
3909       This fixes a long-standing bug where types and even dynamic values
3910       can get internally confused, thereby compromising type safety
3911       (abstraction) and dynamic correctness.  See
3912    
3913         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3914    
3915       for an example of how things could go wrong until now.
3916    
3917       The downside of the new scheme is that pids are not quite as
3918       persistent as they used to be: CM will generate a fresh pid
3919       for every compilation unit that it thinks it sees for the first
3920       time.  That means that if you compile starting from a clean, fresh
3921       source tree at two different times, you end up with different
3922       binaries.
3923    
3924       Cutoff recompilation, however, has not been compromised because
3925       CM keeps pid information in special caches between runs.
3926    
3927    ----------------------------------------------------------------------
3928    Name: Lal George
3929    Date: 2002/02/07 15:34:13 EST 2002
3930    Tag: <none>
3931    Description:
3932    
3933    Compilers that generate assembly code may produce  global labels
3934    whose value is resolved at link time. The various peephole optimization
3935    modules did not take this in account.
3936    
3937    TODO. The Labels.addrOf function should really return an option
3938    type so that clients are forced to deal with this issue, rather
3939    than an exception being raised.
3940    
3941    ----------------------------------------------------------------------
3942    Name: Lal George
3943    Date: 2002/02/06 13:55:02 EST
3944    Tag: george-20020206-ra-breakup
3945    Description:
3946    
3947    1. A bug fix from Allen.
3948    
3949        A typo causes extra fstp %st(0)'s to be generated at compensation
3950        edges, which might cause stack underflow traps at runtime.  This
3951        occurs in fft where there are extraneous fstps right before the 'into'
3952        trap instruction (in this case they are harmless since none of the
3953        integers overflow.)
3954    
3955    2. Pulled out various utility modules that were embedded in the modules
3956       of the register allocator. I need these modules for other purposes, but
3957       they are not complete enough to put into a library (just yet).
3958    ----------------------------------------------------------------------
3959    Name: Matthias Blume
3960    Date: 2002/01/31 16:05:00 EST
3961    Tag: blume-20020131-sparc-ccalls
3962    Description:
3963    
3964    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3965       of extra stack space by mistake.  Fixed.
3966    
3967    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3968    
3969    ----------------------------------------------------------------------
3970    Name: Allen Leung
3971    Date: 2002/01/30
3972    Tag: leunga-20020130-nowhere-bug-fix
3973    Description:
3974    
3975       MLRISC bug fixes:
3976       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3977       2. MachineInt.fromString was negating its value.
3978    
3979    ----------------------------------------------------------------------
3980    Name: Matthias Blume
3981    Date: 2002/01/29
3982    Tag: blume-20020129-INSTALL
3983    Description:
3984    
3985    - Added somewhat detailed installation instructions (file INSTALL).
3986    - Fixed curl-detection bug in config/install.sh.
3987    - It is now possible to select the URL getter using the URLGETTER
3988      environment variable:
3989    
3990          not set / "unknown"      --> automatic detection (script tries wget,
3991                                       curl, and lynx)
3992          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3993                                       how to properly invoke them)
3994          other                    --> use $URLGETTER directly, it must take
3995                                       precisely two command-line arguments
3996                                       (source URL and destination file name)
3997    
3998    ----------------------------------------------------------------------
3999    Name: Matthias Blume
4000    Date: 2002/01/28
4001    Tag: blume-20020128-sparc-ccalls
4002    Description:
4003    
4004    - Fixed problem with calculation of "used" registers in sparc-c-calls.
4005    - Make use of the allocParam argument in sparc-c-calls.
4006    
4007    ----------------------------------------------------------------------
4008    Name: Matthias Blume
4009    Date: 2002/01/28
4010    Tag: blume-20020128-allocParam
4011    Description:
4012    
4013    John Reppy:  Changes c-calls API to accept client-callback for
4014    allocating extra stack space.
4015    me: Corresponding changes to mlriscGen (using a dummy argument that
4016        does not change the current behavior).
4017    
4018    ----------------------------------------------------------------------
4019    Name: Matthias Blume
4020    Date: 2002/01/28 12:00:00
4021    Tag: Release_110_38
4022    Description:
4023    
4024    This time for real!!!
4025    
4026    ----------------------------------------------------------------------
4027    Name: Matthias Blume
4028    Date: 2002/01/28 10:56:00 EST
4029    Tag: blume-20020128-retraction
4030    Description:
4031    
4032    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
4033       with blume-Release_110_38-retracted.)
4034    
4035    1. Fixed a problem with incorrect rounding modes in real64.sml.
4036       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
4037    
4038    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
4039       and enums fixed.  The naming of corresponding ML identifiers should
4040       now be consistent again.
4041    
4042    ----------------------------------------------------------------------
4043    Name: Allen Leung
4044    Date: 2002/01/27
4045    Tag: leunga-20020127-nowhere
4046    Description:
4047    
4048       Added a target called nowhere in the configuration scripts.
4049       Enabling this will build the MLRISC 'nowhere' tool (for translating
4050       programs with where-clauses into legal SML code) during installation.
4051    
4052    ----------------------------------------------------------------------
4053    Name: Matthias Blume
4054    Date: 2002/01/25 21:27:00 EST
4055    Tag: blume-Release_110_38-retracted
4056    Description:
4057    
4058    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
4059    
4060    README will be added later.
4061    
4062    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
4063    (Release_110_38) removed.  Reason: Last-minute bug fixes.
4064    
4065    ----------------------------------------------------------------------
4066    Name: Matthias Blume
4067    Date: 2002/01/25
4068    Tag: blume-20020125-ffi
4069    Description:
4070    
4071    A large number of tweaks and improvements to ml-nlffi-lib and
4072    ml-nlffigen:
4073    
4074       - ML represenation types have been streamlined
4075       - getter and setter functions work with concrete values, not abstract
4076         ones where possible
4077       - ml-nlffigen command line more flexible (see README file there)
4078       - some bugs have been fixed (hopefully)
4079    
4080    ----------------------------------------------------------------------
4081    Name: Lal George
4082    Date: 2002/01/24
4083    Tag: george-20020124-risc-ra-interface
4084    Description:
4085    
4086       There is a dramatic simplification in the interface to the
4087       register allocator for RISC architectures as a result of making
4088       parallel copy instructions explicit.
4089    
4090    ----------------------------------------------------------------------
4091    Name: Matthias Blume
4092    Date: 2002/01/22
4093    Tag: blume-20020122-x86-ccalls
4094    Description:
4095    
4096    Bug fix for c-calls on x86 (having to do with how char- and
4097    short-arguments are being handled).
4098    
4099    ----------------------------------------------------------------------