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