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/trunk/NOTES/HISTORY
ViewVC logotype

Diff of /sml/trunk/NOTES/HISTORY

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

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