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/HISTORY
ViewVC logotype

Diff of /sml/trunk/HISTORY

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

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