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 800, Fri Mar 16 17:22:47 2001 UTC revision 1751, Mon Jan 24 23:39:13 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: 2005/01/24 17:40:00 CST
18    Tag: blume-20050124-mlyacc
19    Description:
20    
21    Minor cleanup in ML-Yacc rule printing mechanism.  This should fix a
22    problem with certain "as" patterns which previously got rendered
23    using incorrect syntax.
24    
25    ----------------------------------------------------------------------
26    Name: Matthias Blume (blume (at) tti - c (dot) org)
27    Date: 2005/01/18 12:00:00 CST
28    Tag: blume-20050118-profile
29    Description:
30    
31    Made time profiling code (interrupt handler) in runtime system aware
32    of new array representation.
33    
34    ----------------------------------------------------------------------
35    Name: Matthias Blume (blume (at) tti - c (dot) org)
36    Date: 2005/01/14 18:00:00 CST
37    Tag: blume-20050114-heap2exec
38    Description:
39    
40    Implemented new (but still experimental) heap2exec facility.  This is
41    tested under Mac OS X and should work under Linux (will test shortly).
42    It will probably also work on the Sparc (will test some time later).
43    
44      - removed old "HACKED_STANDALONE" hack from runtime
45    
46    To be able to test this, uncomment the request for "heap2asm" in
47    config/targets prior to installation.  (Notice that this is different
48    from "heap2exec" mentioned below.  Not a typo.)
49    
50    To perform an actual test, run the command
51    
52       $ bin/heap2exec heapfile execfile
53    
54    (You can put heap2exec on your shell's path.)
55    
56    For example, run
57    
58      $ bin/heap2exec bin/.heap/ml-yacc.ppc-darwin mly
59    
60    This will create a standalone executable called "mly" which you can
61    then invoke directly as a command.
62    
63    ----------------------------------------------------------------------
64    Name: Matthias Blume (blume (at) tti - c (dot) org)
65    Date: 2005/01/07 11:44:00 CST
66    Tag: blume-20050107-mlstring
67    Description:
68    
69    fixed off-by-one error in ML_STRING macro (globals.c)
70    
71    ----------------------------------------------------------------------
72    Name: Matthias Blume (blume (at) tti - c (dot) org)
73    Date: 2004/12/23 18:00:00 CST
74    Tag: blume-20041223-santa
75    Description:
76    
77    Made ml-build script "smarter" (but only very little).
78    
79    ----------------------------------------------------------------------
80    Name: Matthias Blume (blume (at) tti - c (dot) org)
81    Date: 2004/12/21 15:05:00 CST
82    Tag: blume-20041221-longlong
83    Description:
84    
85    * Implemented access to signed and unsigned long long data in NLFFI.
86    (The parameter-passing part of the picture has not complete. But
87    data structure access seems to work.)
88    
89    * Fixed CM's incorrect assumption that the PPC is little-endian.
90      (On the Mac, it is big-endian.  And that's currently our only
91      PPC platform.)
92    
93    ----------------------------------------------------------------------
94    Name: Matthias Blume (blume (at) tti - c (dot) org)
95    Date: 2004/12/21 12:50:00 CST
96    Tag: blume-20041221-memory
97    Description:
98    
99    Some cleanup in the $c/memory.cm library: separated some concerns by
100    moving allocation code and memory access code each into their own
101    files.
102    
103    ----------------------------------------------------------------------
104    Name: Allen Leung (leunga (at) reservoir (dot) com)
105    Date: 2004/12/17 16:12:00 EST
106    Tag: leunga-20041217-cygwin-smlnj-home
107    Description:
108    
109        The Unix I/O library of SML/NJ on cygwin does not understand
110    Windows style pathname, so problems arise when SMLNJ_HOME is set
111    to a Windows style pathname.  _run-sml now converts SMLNJ_HOME
112    to a POSIX pathname on cygwin.
113    
114    ----------------------------------------------------------------------
115    Name: Matthias Blume (blume (at) tti - c (dot) org)
116    Date: 2004/12/16 13:35:00 CST
117    Tag: Release_110_52
118    Description:
119    
120    Last-minute changes incorporated into 110.52.  Release tag moved.
121    
122    The changes:
123       - HashString.hashString' -> HashString.hashSubstring
124       - bug fix in UnivariateStats
125    
126    ----------------------------------------------------------------------
127    Name: Matthias Blume (blume (at) tti - c (dot) org)
128    Date: 2004/12/15 23:40:00 CST
129    Tag: blume-20041215-hashSubstring
130    Description:
131    
132    - HashString.hashString' -> HashString.hashSubstring
133    - corresponding changes in atom.sml
134    - "de-compressed" (aka. un-obfuscated) code for UnivariateStats and
135      added some comments
136    
137    ----------------------------------------------------------------------
138    Name: Matthias Blume (blume (at) tti - c (dot) org)
139    Date: 2004/12/15 15:30:00 CST
140    Tag: (Release_110_52)
141    Description:
142    
143    New working version (110.52).  NEW BOOTFILES!
144    
145    ----------------------------------------------------------------------
146    Name: Matthias Blume (blume (at) tti - c (dot) org)
147    Date: 2004/12/15 12:45:00 CST
148    Tag: blume-20041215-spaces
149    Description:
150    
151    More on the space problem (this time for Win32).
152    
153    ----------------------------------------------------------------------
154    Name: Matthias Blume (blume (at) tti - c (dot) org)
155    Date: 2004/12/14 17:30:00 CST
156    Tag: blume-20041214-spaces
157    Description:
158    
159    Hacked some of the scripts (in particular: the installer) to cope with
160    spaces in filenames a bit better.  But beware: the current "solution"
161    is likely still full of bugs and inherently incomplete.  (We need to
162    do away with those shell scripts for a comprehensive solution.)
163    
164    ----------------------------------------------------------------------
165    Name: Matthias Blume (blume (at) tti - c (dot) org)
166    Date: 2004/12/13 14:45:00 CST
167    Tag: blume-20041213-ml-makedepend
168    Description:
169    
170    Fixed bug in code for ml-makedepend.
171    
172    ----------------------------------------------------------------------
173    Name: Matthias Blume (blume (at) tti - c (dot) org)
174    Date: 2004/12/09 16:30:00 CST
175    Tag: blume-20041209-statistics
176    Description:
177    
178    Added two simple but potentially useful statistics modules to SML/NJ Library.
179    (See CHANGES file there.)
180    
181    ----------------------------------------------------------------------
182    Name: Matthias Blume (blume (at) tti - c (dot) org)
183    Date: 2004/12/01 16:50:00 CST
184    Tag: blume-20041201-atom
185    Description:
186    
187    smlnj-lib:
188    
189      Added function HashString.hashString' for substrings.
190      Hand-inlined CharVector.foldl into HashString (for speed).
191      Modified implementation of structure Atom to avoid extracting
192      strings from substrings unless necessary.
193      (Also see CHANGES file for smlnj-lib.)
194    
195    ----------------------------------------------------------------------
196    Name: Matthias Blume (blume (at) tti - c (dot) org)
197    Date: 2004/11/24 22:15:00 CST
198    Tag: blume-20041124-cml
199    Description:
200    
201    Made sure CML compiles when Position = Int64.
202    
203    ----------------------------------------------------------------------
204    Name: Matthias Blume (blume (at) tti - c (dot) org)
205    Date: 2004/11/24 14:30:00 CST
206    Tag: blume-20041124-position
207    Description:
208    
209    The compiler can now be compiled in a mode that makes structure Position
210    equal to Int64.  The default, however, is unchanged (Position = Int31)
211    for the time being.
212    
213    To enable 64-bit positions, use the following procedure:
214    
215    1. Start sml
216    2. Autoload $smlnj/cmb.cm (if not already autoloaded)
217    3. Type
218          #set (CMB.symval "USE_64_BIT_POSITIONS") (SOME 1);
219    4. Run CMB.make() as usual.
220    
221    This is barely tested.  The only test so far was a little SML program
222    counting the number of characters in an 8-gigabyte file by
223    reading it character-by-character.  That test was successful.
224    
225    In support of 64-bit positions, a number of new functions have been
226    added to the runtime system.
227    
228    ----------------------------------------------------------------------
229    Name: Matthias Blume (blume (at) tti - c (dot) org)
230    Date: 2004/11/23 14:45:00 CST
231    Tag: blume-20041123-useFile
232    Description:
233    
234    Fixed a problem with unhelpful error messages related to problems with
235    .cm- or .sml files that appear as part of the sml command line.
236    
237    ----------------------------------------------------------------------
238    Name: Matthias Blume (blume (at) tti - c (dot) org)
239    Date: 2004/11/18 15:40:00 CST
240    Tag: Release_110_51
241    Description:
242    
243    New working version (110.51).  NEW BOOTFILES!
244    
245    ----------------------------------------------------------------------
246    Name: Matthias Blume (blume (at) tti - c (dot) org)
247    Date: 2004/11/18 15:35:00 CST
248    Tag: <none>
249    Description:
250    
251    Enabled dlopen and friends for FreeBSD (as recommended by Johannes 5
252    Joemann).
253    
254    ----------------------------------------------------------------------
255    Name: Allen Leung (leunga (at) reservoir (dot) com)
256    Date: 2004/11/17 16:05:21 EST 2004
257    Tag: leunga-20041117-mlrisc-live-kill
258    Description:
259    
260       Added support for MLTree constructs LIVE and KILL
261    to all the architectures.
262    
263    ----------------------------------------------------------------------
264    Name: Matthias Blume (blume (at) tti - c (dot) org)
265    Date: 2004/11/13 00:20:00 CST
266    Tag: blume-20041113-versiontool
267    Description:
268    
269    - Stripped down the versiontool: It now only handles the version number.
270      The date string is generated at bootstrap time (during makeml).
271    
272    - In a previous commit, fixed a minor issue with how polyequal is being
273      translated.  In particular, the code now "looks through" abstractions.
274      This results in slightly fewer polyEqual warnings and hopefully slightly
275      more efficient code.  Important examples for where this matters are
276      the new int64 and word64 types.
277    
278    ----------------------------------------------------------------------
279    Name: Matthias Blume (blume (at) tti - c (dot) org)
280    Date: 2004/11/12 00:30:00 CST
281    Tag: blume-20041112-int64
282    Description:
283    
284    Structure Int64 fully hooked in.  (The implementation is not very
285    efficient, though.)
286    
287    ----------------------------------------------------------------------
288    Name: Matthias Blume (blume (at) tti - c (dot) org)
289    Date: 2004/11/11 17:30:00 CST
290    Tag: blume-20041111-more64
291    Description:
292    
293    All the pieces of Word64 are now there, with the exception of the
294    conversions from and to LargeWord. (Eventually these need to be identities,
295    but for the time being they don't even make sense because LargeWord is
296    32-bit wide.)
297    
298    Also started to add similar support for Int64, but major pieces of that
299    are still missing.
300    
301    ----------------------------------------------------------------------
302    Name: Matthias Blume (blume (at) tti - c (dot) org)
303    Date: 2004/11/11 00:15:00 CST
304    Tag: blume-20041111-word64
305    Description:
306    
307    Structure Word64 is now (almost) complete, word literals and patterns
308    seem to work.  There are a few odd pieces missing.  In particular,
309    I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
310    at the moment.
311    
312    Making Word64 official would mean that LargeWord becomes Word64.  But
313    this requires extreme care because most word-word conversions have to
314    go through LargeWord, so making a mistake means loss of efficiency or
315    worse.  Eventually there will be a solution similar to (but actually
316    simpler than) what I did with IntInf.
317    
318    ----------------------------------------------------------------------
319    Name: Matthias Blume (blume (at) tti - c (dot) org)
320    Date: 2004/11/10 18:12:00 CST
321    Tag: blume-20041110-64bit
322    Description:
323    
324    More 64-bit hacking (but still not even half-way there yet).
325    Also, some assorted improvements to the handling of 8-bit words.
326    
327    ----------------------------------------------------------------------
328    Name: Matthias Blume (blume (at) tti - c (dot) org)
329    Date: 2004/11/09 17:50:00 CST
330    Tag: <none>
331    Description:
332    
333    Started adding some infrastructure for supporting 64-bit int- and
334    word-types.  (Still in its very early stages.)
335    
336    ----------------------------------------------------------------------
337    Name: Matthias Blume (blume (at) tti - c (dot) org)
338    Date: 2004/10/28 10:45:00 CDT
339    Tag: Release_110_50
340    Description:
341    
342    New working version (110.50).  NEW BOOTFILES!
343    
344                 =====================
345    
346    Also:
347    
348      - Changend config/srcarchiveurl from a file just
349        containing the URL string into a file containing
350        shell script code.  The code has access to the $VERSION variable.
351      - Made corresponding changes to config/install.sh and config/unpack.
352      - Default contents of config/srcarchiveurl uses $VERSION and
353        normally does not have to be edited to reflect a version change.
354    
355      (As a result, a version change can be done by just editing
356       config/version, the rest is now automatic.)
357    
358    ----------------------------------------------------------------------
359    Name: Matthias Blume (blume (at) tti - c (dot) org)
360    Date: 2004/10/27 17:50:00 CDT
361    Tag: blume-20041027-btrace-msg
362    Description:
363    
364    BackTrace.monitor now also reports the source of the exception that
365    triggered the trace.
366    
367    ----------------------------------------------------------------------
368    Name: Matthias Blume (blume (at) tti - c (dot) org)
369    Date: 2004/10/27 17:20:00 CDT
370    Tag: blume-20041027-x86-c-calls
371    Description:
372    
373    This is the HISTORY entry for two earlier commits, both concerning
374    the x86 c-calls code in MLRISC:
375    
376      - added a missing LOAD in the code that deals with struct arguments
377      - made sure the caller does not add the wrong number of bytes to the
378        stack pointer after a call of a function returning a struct
379        (the callee already pops the implicit argument which points to
380         the space reserved for the result)
381    
382    ----------------------------------------------------------------------
383    Name: Allen Leung (leunga (at) reservoir (dot) com)
384    Date: 2004/10/24 14:00:00 EST
385    Tag: leunga-20041024-x86-gas-fucomip
386    Description:
387    
388       John discovered a bug in the syntax of fucomip.
389       The opcodes FU?COMIP? have been changed to
390    
391           fu?comip? %st(i), %st
392    
393    ----------------------------------------------------------------------
394    Name: Matthias Blume (blume (at) tti - c (dot) org)
395    Date: 2004/10/20 15:06:00 CDT
396    Tag: blume-20041020-standalone-backtrace
397    Description:
398    
399    Added a mechanism for getting back-trace information from standalone
400    programs.  Here is how it works:
401    
402       1. The part of the program from which you want to get backtrace
403          information (usually the whole program) should be wrapped with
404          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
405          main program could be modified from something like
406    
407             fun main (pgm, args) = ...
408    
409         to
410    
411            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
412    
413       2. To be able to access BackTrace.monitor, you have to add
414          library $smlnj-tdp/plugins.cm to the .cm file that contains your
415          main function.
416    
417       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
418          CM might have created in the past for your project).
419    
420       4. Build the system using this command line:
421    
422             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
423                      myprog.cm MyProg.main myprog
424    
425          instead of the usual
426    
427             ml-build myprog.cm MyProg.main myprog
428    
429    I changed a library name:
430    
431       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
432    
433    New libraries:
434    
435       $smlnj-tdp/back-trace.cm
436          -- when loaded causes the back-trace plugin to be installed
437       $smlnj-tdp/coverage.cm
438          -- when loaded causes the coverage plugin to be installed
439    
440    ----------------------------------------------------------------------
441    Name: Matthias Blume (blume (at) tti - c (dot) org)
442    Date: 2004/10/18 16:45:00 CDT
443    Tag: blume-20041018-groupowner
444    Description:
445    
446    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
447    
448    Eliminated group owner specs from .cm files throughout the source tree.
449    
450    ----------------------------------------------------------------------
451    Name: Matthias Blume (blume (at) tti - c (dot) org)
452    Date: 2004/10/15 15:45:00 CDT
453    Tag: blume-20041015-coverage
454    Description:
455    
456    * Test coverage tool added!
457    * Further reorganization of tracing-, debugging-, and profiling support:
458    
459        - moved original BTImp -- now called BackTrace -- into a separate
460          library called $/trace-debug-profile.cm
461        - eliminated all mentions of BTrace from SMLofNJ.Internals
462        - only the instrumentation mechanism is now left in the compiler proper
463        - BackTrace module is a plugin which is NOT plugged in by default
464        - Coverage module is another such plugin
465    
466    To get the benefits of any of these plugin modules, the code in
467    question must be compiled with tdp instrumentation turned on.  This
468    can be done by setting SMLofNJ.Internals.TDP.mode to true.
469    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
470    
471    Plugins are selected at link time.  (Pre-compiled instrumented code
472    can be re-loaded with different plugins in effect.)  When an
473    instrumented module is linked, whatever plugins are at that time
474    enabled will come into effect for that module.
475    
476    To enable the back-trace plugin, load library $/trace-debug-profile.cm
477    and invoke BackTrace.install() (e.g., from the interactive prompt).
478    To enable the coverage plugin, load the same library and invoke
479    Coverage.install().
480    
481    Back-traces are generated automatically on uncaught exceptions and
482    when the code in question explicitly invokes BackTrace.trigger().
483    
484    Coverage (and execution frequency-) information must be queried
485    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
486    
487    ----------------------------------------------------------------------
488    Name: Matthias Blume (blume (at) tti - c (dot) org)
489    Date: 2004/10/14 17:40:00 CDT
490    Tag: blume-20041014-tdp-core
491    Description:
492    
493    Snapshot of a significant overhaul of how the trace/debug/profile support
494    is hooked into the system (specifically: Core and SMLofNJ.Internals).
495    
496    ----------------------------------------------------------------------
497    Name: Matthias Blume (blume (at) tti - c (dot) org)
498    Date: 2004/10/13 16:34:00 CDT
499    Tag: blume-20041013-tdp
500    Description:
501    
502    Some rationalization of names:
503    
504       structure BTrace -> structure TDPInstrument
505       etc.
506    
507    This is is preparation of using the original back-trace
508    instrumentation for other purposes.  "TDP" stands for
509    Trace/Debug/Profile.
510    
511    The control flag controlling whether instrumentation is on or off is now
512    registered under a different name, so instead of running sml as
513    
514     sml -Cinstrument.btrace-mode=true
515    
516    one has to say
517    
518     sml -Ctdp.instrument=true
519    
520    ----------------------------------------------------------------------
521    Name: Matthias Blume (blume (at) tti - c (dot) org)
522    Date: 2004/10/11 16:37:00 CDT
523    Tag: blume-20041011-regions
524    Description:
525    
526    Made some minor modifications to elabcore.sml to have source regions
527    be propagated more tightly -- resulting in better (i.e., smaller)
528    regions being reported in error- and debug messages.
529    
530    ----------------------------------------------------------------------
531    Name: Matthias Blume (blume (at) tti - c (dot) org)
532    Date: 2004/10/08 22:50:00 CDT
533    Tag: blume-20041008-cmkw
534    Description:
535    
536    Fixed handling of keywords in .cm files: After seeing "is" the lexer
537    treats subsequent occurrences of "group", "library", "source", "is",
538    "*", and "-" as ordinary identifiers rather than keywords.
539    
540    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
541    is supposed to accept a tool argument called "source", but this did
542    not work because of the clash with the keyword.
543    
544    ----------------------------------------------------------------------
545    Name: Matthias Blume (blume (at) tti - c (dot) org)
546    Date: 2004/10/07 16:00:00 CDT
547    Tag: blume-20041007-cleanup
548    Description:
549    
550    Assorted cleanup work:
551    
552      - got rid of intstrmap in favor of using the library's
553        hash table implementation
554      - threw out most of the pathnames stuff, as it was not used anyway
555      - simplified tokentable implementation
556      - fixed some minor spelling errors
557    
558    ----------------------------------------------------------------------
559    Name: Matthias Blume (blume (at) tti - c (dot) org)
560    Date: 2004/10/06 15:15:15 CDT
561    Tag: blume-20041006-handler
562    Description:
563    
564    Cleaned up the absyn to reflect the invariant that HANDLE always
565    carries a FNexp as part of the type definition.  This eliminates some
566    superfluous sanity checks at runtime down the road.
567    
568    Some minor cleanup of the btrace code.
569    
570    ----------------------------------------------------------------------
571    Name: Matthias Blume (blume (at) tti - c (dot) org)
572    Date: 2004/10/01 10:20:30 CDT
573    Tag: blume-20041001-slave
574    Description:
575    
576    Added hack to make slave mode work in the presence of the version
577    tool.  (Still, since the master does two passes over the code for
578    CMB.make, the release number gets bumped twice when slaves are
579    attached.  I don't know if this is worth fixing...)
580    
581    ----------------------------------------------------------------------
582    Name: Matthias Blume (blume (at) tti - c (dot) org)
583    Date: 2004/09/30 10:55:00 CDT
584    Tag: blume-20040930-version
585    Description:
586    
587    * Moved the "version" magic into its own little library under
588      src/system/smlnj/internal.  This avoids expensive reconstruction of
589      a stable src/compiler/core.cm.
590    
591    * At the same time, structure CompilerVersion is now known as
592      structure SMLNJVersion.
593    
594    * Arranged for the version tool to NOT kick in when rebuilding the system
595      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
596      Also, loading the versiontool does not work when rebuilding the system
597      because CM is not properly initialized at that time.
598    
599    ----------------------------------------------------------------------
600    Name: Matthias Blume (blume (at) tti - c (dot) org)
601    Date: 2004/09/29 14:00:00 CDT
602    Tag: blume-20040929-autoversion
603    Description:
604    
605    Implemented some CM magic to have
606        file src/compiler/TopLevel/main/version.sml
607    generated automagically.
608    The version is taken from two files: config/version and config/release.
609    The first is expected to contain a two-part version number such as 110.49.
610    The second should contain a single number, but it may be missing.
611    
612    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
613    time the version tool is loaded (which is the first time you say CMB.make),
614    then the tool will increment the value stored in config/release every
615    time CMB.make is invoked.
616    
617    The binfile format is now insensitive to anything beyond the first
618    two components of a version number, so bumping the release does not render
619    binfiles incompatible.  Auto-bumping can be used to keep track of versions
620    during development without invalidating existing binfiles.
621    
622    In any case, every CMB.make updates the date information in version.sml.
623    (This is the date that is printed in the banner.)
624    
625    ----------------------------------------------------------------------
626    Name: Matthias Blume (blume (at) tti - c (dot) org)
627    Date: 2004/09/28 10:53:00 CDT
628    Tag: blume-20040928-controls
629    Description:
630    
631    Some cleanup of the controls code.
632    
633    ----------------------------------------------------------------------
634    Name: Matthias Blume (blume (at) tti - c (dot) org)
635    Date: 2004/09/27 22:08:00 CDT
636    Tag: blume-20040927-controls
637    Description:
638    
639    Added two pieces of functionality to the Controls interface:
640    
641    1.   val save'restore: 'a control -> unit -> unit
642    
643       grabs the current value of the control in stage 1 and restores it
644       in stage 2.
645    
646    2.   val set' : 'a control * 'a -> unit -> unit
647    
648       stores the given value into the control in stage 2 (i.e., delayed)
649       but does all error checking in stage 1.
650       (This is for string controls that need to do parse their argument --
651        something that might fail.  In some cases, notably in CM, one
652        already knows the intended argument but wants to delay the actual
653        assignment until a time when error recovery would be more difficult.)
654    
655    Changed the handling of controls in tool arguments to classes "sml" and
656    "lazysml":
657       - use Controls.save'restore as a more robust way of restoring the
658         old value (in particular: without having to re-parse the string)
659       - use controls to handle the "overload" keyword in the init group
660         (I believe this change actually fixes a long-standing obscure bug.)
661    
662    ----------------------------------------------------------------------
663    Name: Matthias Blume (blume (at) tti - c (dot) org)
664    Date: 2004/09/27 17:00:00 CDT
665    Tag: blume-20040927-lazysml
666    Description:
667    
668    Added a new tool class called "lazysml" to CM's tool chest.  The only
669    difference to "sml" is that compilation is done with Control.lazysml
670    set to true.  A source of class "lazysml" is automatically recognized
671    by a file name suffix of ".lml".
672    
673    In addition to the above feature, the original class "sml" now also
674    supports a tool argument "lazy" which has the same effect.  As a
675    result, the following three lines are equivalent:
676    
677        foo.sml : lazysml
678        foo.sml : sml (lazy)
679        foo.sml (lazy)
680    
681    The setting goes into effect both during parsing and during
682    compilation.  The original setting is restored right after parsing and
683    after compilation, respectively.
684    
685    In addition to all the above, there is also a general mechanism to set
686    ANY of the "controls" that are available at the command line via
687    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
688    apply as well. (In fact, "lazy" is implemented as a special case of
689    the general mechanism.)
690    
691    The .cm file syntax uses a new keyword tool argument called "with".
692    There are several ways of indicating the desired settings:
693    
694        foo.sml (with:parser.quotations=true)
695        foo.sml (with:(name:parser.quotations value:true))
696        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
697        foo.sml (with:(name1=value1 name2=value2 ...))
698        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
699    
700        etc.
701    
702    Another possible abbreviation is to leave out the =v or value:v part
703    if the name refers to a boolean control (in which case the value is
704    taken to be true).  Thus, one could get lazy sml also by saying:
705    
706        foo.sml (with:parser.lazy-keyword=true)
707        foo.sml (with:parser.lazy-keyword)
708        foo.sml (with:(name:parser.lazy-keyword value:true))
709        foo.sml (with:(name:parser.lazy-keyword))
710    
711    ----------------------------------------------------------------------
712    Name: Matthias Blume (blume (at) tti - c (dot) org)
713    Date: 2004/09/24 16:22:00 CDT
714    Tag: blume-20040924-ppc-long-branch
715    Description:
716    
717    Turned message about "emiting long form of branch" off by default.
718    Added a control flag to turn it back on when desired.
719    
720    ----------------------------------------------------------------------
721    Name: Matthias Blume (blume (at) tti - c (dot) org)
722    Date: 2004/09/24 16:05:00 CDT
723    Tag: blume-20040924-rounding
724    Description:
725    
726    Applied patch for setting rounding modes under Mac OS X.  Thanks to
727    Melissa O'Neill for providing the code!
728    
729    ----------------------------------------------------------------------
730    Name: Matthias Blume (blume (at) tti - c (dot) org)
731    Date: 2004/09/23 17:30:00 CDT
732    Tag: blume-20040923-envvars
733    Description:
734    
735    1. Changed definition of type ControlRegistry.registry_tree to
736       include control_info (i.e., the name of the controlling environment
737       variable).
738    
739    2. Added command-line flags -e and -E to print the names of environment
740       variables that can be used to control internal settings.
741       (This uses the new API mentioned in 1.)
742    
743    ----------------------------------------------------------------------
744    Name: Matthias Blume (blume (at) tti - c (dot) org)
745    Date: 2004/09/13 16:50:00 CDT
746    Tag: Release_110_49
747    Description:
748    
749    New working version (110.49).  NEW BOOTFILES!
750    
751    ----------------------------------------------------------------------
752    Name: Matthias Blume
753    Date: 2004/09/13 16:20:00 CDT
754    Tag: blume-20040913-config-mlrisc
755    Description:
756    
757    Put target "mlrisc" back into the default list.
758    (There is no harm in having it, and some users have expressed their
759    wish to have "mlrisc" included by default.)
760    
761    ----------------------------------------------------------------------
762    Name: John Reppy
763    Date: 2004/09/13
764    Tag: jhr-20040913-signals
765    Description:
766    
767    Fixed the signal masking code to properly nest mask/unmask operations
768    on a per-signal basis.
769    
770    ----------------------------------------------------------------------
771    Name: Matthias Blume (blume (at) tti - c (dot) org)
772    Date: 2004/09/08 13:20:00 CDT
773    Tag: blume-20040908-heap-magic
774    Description:
775    
776    Bumped the heap macig number to 0x09082004 to account for the changed
777    layout of the ML frame under MacOS X.
778    
779    ----------------------------------------------------------------------
780    Name: Allen Leung (leunga (at) reservoir (dot) com)
781    Date: 2004/09/03 11:26:00 EST
782    Tag: leunga-20040903-cygwin-install
783    Description:
784    
785       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
786    The Cygwin runtime is turned on by setting the environment
787    variable SMLNJ_CYGWIN_RUNTIME to 1.
788    
789    ----------------------------------------------------------------------
790    Name: Matthias Blume (blume (at) tti - c (dot) org)
791    Date: 2004/08/31 17:15:00 CDT
792    Tag: blume-20040831-core
793    Description:
794    
795    Added some exports to src/compiler/core.cm upon request by J. Joemann.
796    
797    ----------------------------------------------------------------------
798    Name: Matthias Blume (blume (at) tti - c (dot) org)
799    Date: 2004/08/30 17:55:00 CDT
800    Tag: blume-20040830-installer
801    Description:
802    
803    Upon request by Johannes Joemann:
804    
805    - improved ML code of installer to fall back to coping when renaming
806      fails (i.e., when source and target are on different file systems);
807      the code compiles but has yet to be tested in anger
808    - removed mlrisc from list of default targets (config/targets)
809    
810    ----------------------------------------------------------------------
811    Name: Matthias Blume (blume (at) tti - c (dot) org)
812    Date: 2004/08/27 17:20:00 CDT
813    Tag: blume-20040827-ptreql
814    Description:
815    
816    Added ptreql primop to structure InlineT (upon request from Larry
817    Paulson).
818    
819    ----------------------------------------------------------------------
820    Name: Allen Leung (leunga (at) reservoir (dot) com)
821    Date: 2004/08/15 21:21:00 EST
822    Tag: leunga-110_48-udgraph
823    Description:
824    
825    Another bug fix from Carl Hauser:
826    
827    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
828    > 48c48
829    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
830    > e::L)
831    > ---
832    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
833    > Without this, any deletion of an edge in an undirected graph does severe
834    > violence to the graph.
835    
836    ----------------------------------------------------------------------
837    Name: Allen Leung (leunga (at) reservoir (dot) com)
838    Date: 2004/08/10 23:35:00 EST
839    Tag: leunga-110_48-ppc
840    Description:
841    
842       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
843    
844    ----------------------------------------------------------------------
845    Name: Matthias Blume (blume (at) tti - c (dot) org)
846    Date: 2004/08/10 12:00:00 CDT
847    Tag: Release_110_48
848    Description:
849    
850    New working version (110.48).  NEW BOOTFILES!
851    
852    ----------------------------------------------------------------------
853    Name: Allen Leung (leunga (at) reservoir (dot) com)
854    Date: 2004/08/09 12:21:00 EST
855    Tag: leunga-110_47-dijsktra
856    Description:
857    
858    Bug fix from Carl Hauser:
859    
860    single_source_shortest_paths in dijkstra.sml was observed to get wrong
861    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
862    
863    The problem is that following the expression A.update(dist,s,Num.zero)
864    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
865    
866    ----------------------------------------------------------------------
867    Name: Matthias Blume (blume (at) tti - c (dot) org)
868    Date: 2004/08/06 18:10:00 CDT
869    Tag: blume-20040806-cmdline
870    Description:
871    
872    Fiddled with handling of command-line options:
873    
874      * sml now quits after processing the command line
875        if -H, -S, -h<n>, or -s<n> appears as the last
876        command-line argument
877      * a new option -q terminates the session when encountered on
878        the command line; subsequent arguments will be ignored
879      * bug fixes: short (erroneous) arguments are no longer ignored
880        completely
881    
882    ----------------------------------------------------------------------
883    Name: Allen Leung (leunga (at) reservoir (dot) com)
884    Date: 2004/08/04 18:17:00 EST
885    Tag: leunga-110_47-ppc-ibm-asm
886    Description:
887    
888      - Added minimal IBM assembly syntax support for PowerPC.
889    
890      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
891        been altered in the runtime.  We need an automatic way to keep the file
892        in sync.
893    
894    ----------------------------------------------------------------------
895    Name: Matthias Blume (blume (at) tti - c (dot) org)
896    Date: 2004/08/04 14:00:00 CDT
897    Tag: Release_110_47
898    Description:
899    
900    New working version (110.47).  NEW BOOTFILES!
901    
902    ----------------------------------------------------------------------
903    Name: Matthias Blume (blume (at) tti - c (dot) org)
904    Date: 2004/08/03 14:25:00 CDT
905    Tag: blume-20040803-callingconv
906    Description:
907    
908    Added low-level support for choosing C calling conventions by
909    twiddling the type of rawccall.  (See
910    src/compiler/Semant/types/cproto.sml for details.)
911    
912    ----------------------------------------------------------------------
913    Name: Matthias Blume (blume (at) tti - c (dot) org)
914    Date: 2004/08/02 15:55:00 CDT
915    Tag: blume-20040802-backout
916    Description:
917    
918    Backed out of change to win32-filesys.c.  The earlier patch to
919    get_file_time caused CM to produce files with the wrong time stamp.
920    
921    ----------------------------------------------------------------------
922    Name: Matthias Blume (blume (at) tti - c (dot) org)
923    Date: 2004/08/02 14:45:00 CDT
924    Tag: blume-20040802-nlffi-win32
925    Description:
926    
927    Added NLFFI support for Win32, adapted from a patch provided by David
928    Hansel.  This is currently completely untested.  Also, the issue
929    concerning stdcall vs. ccall is still unresolved.
930    
931    ----------------------------------------------------------------------
932    Name: Matthias Blume (blume (at) tti - c (dot) org)
933    Date: 2004/07/30 17:55:00 CDT
934    Tag: blume-20040730-various
935    Description:
936    
937    Gearing up towards 110.47...
938    
939    - various minor bugfixes to ml-nlffigen
940    - a beginning of a manual for nlffi
941    
942    - eliminated 'export name=value' in config/install.sh as this does
943      not work with certain versions of /bin/sh
944      (Thanks to David King at Motorola for catching this.)
945    
946    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
947      - added a test for tm==NULL to gmtime.c and localtime.c
948      - applied patch for incorrect GetFileTime under win32
949      - toSeconds -> toMilliseconds in Win32/win32-process.sml
950    
951    ----------------------------------------------------------------------
952    Name: Matthias Blume (blume (at) tti - c (dot) org)
953    Date: 2004/07/21 18:20:00 CDT
954    Tag: blume-20040721-nlffigen
955    Description:
956    
957    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
958      for a typedef to an incomplete type, but leave out the "typ" member.
959      (This is just for consistency.)
960    - Started to produce what is supposed to become better (i.e., comprehensive)
961      documentation of what ml-nlffigen does and produces.
962    
963    ----------------------------------------------------------------------
964    Name: Matthias Blume (blume (at) tti - c (dot) org)
965    Date: 2004/07/14 16:25:00 CDT
966    Tag: blume-20040714-union
967    Description:
968    
969    Added C_UNION to c-calls/c-types.sml and updated the machinery
970    (ml-nlffigen, cproto.sml) that conveys C function interface
971    information to the code generator.
972    
973    However, the actual architecture-specific implementation of function
974    arguments and results that are C unions is still not implemented.
975    
976    ----------------------------------------------------------------------
977    Name: Allen Leung (leunga (at) reservoir (dot) com)
978    Date: 2004/07/14 14:38:00 EST
979    Tag: leunga-110_46_1-ppc-lwzu
980    Description:
981    
982        Added these instructions to the PowerPC architecture:
983        LBZU(X), LHZU(X), LWZU(X),
984        STWU(X), STFDU, STFSU
985    
986        etc...
987    
988        Note: I haven't added their instruction encoding into the description.
989    
990    ----------------------------------------------------------------------
991    Name: Allen Leung (leunga (at) reservoir (dot) com)
992    Date: 2004/07/13 15:04:00 EST
993    Tag: leunga-110_46_1-ppc-lwarx
994    Description:
995    
996       Added the two instructions LWARX and STWCX to the PowerPC
997    instruction set.
998    
999       A (untested) rewrite of loop-structure.sml.  The old version
1000    is completely broken.
1001    
1002    ----------------------------------------------------------------------
1003    Name: Matthias Blume (blume (at) tti - c (dot) org)
1004    Date: 2004/07/13 13:50:00 CDT
1005    Tag: blume-20040713-nlffi
1006    Description:
1007    
1008    - use paramAlloc to report c-calls with too many arguments
1009      (for PPC version where parameter area is pre-allocated)
1010    - added ccall_maxargspace to machspec (to implement the above)
1011    - made "make" commend in CM's "make" tool configurable
1012    - added option (default: on) for passing the name of the SML/NJ's "bin"
1013      directory to "make"; the call looks like this:
1014    
1015         make <options> SMLNJ_BINDIR=<dir> <target>
1016    
1017      This can be used by the Makefile to, e.g., pick the "right" version
1018      of ml-nlffigen.
1019    - minor code tweaks
1020    
1021    ----------------------------------------------------------------------
1022    Name: Matthias Blume (blume (at) tti - c (dot) org)
1023    Date: 2004/07/12 22:50:00 CDT
1024    Tag: blume-110_46_1-macosx-nlffi
1025    Description:
1026    
1027    NLFFI under Mac OS X now working (sort of).  This is largely untested,
1028    though.
1029    
1030    Note:  1. You have to make a new, clean build of the runtime system.
1031           2. There are new BOOTFILES, you have to use them!
1032              (Doing the bootstrap process yourself would be *very* painful!
1033               If you absolutely have to do it, build the system under
1034               a different architecture and then cross-compile.)
1035    
1036    Version bumped to 110.46.1 to account for runtime data format changes.
1037    
1038    ----------------------------------------------------------------------
1039    Name: Matthias Blume (blume (at) tti - c (dot) org)
1040    Date: 2004/06/18 14:30:00 CDT
1041    Tag: blume-20040618-unix
1042    Description:
1043    
1044    Changed the implementation of structure Unix so that the same stream
1045    is returned every time one of the {text,bin}{In,Out}streamOf functions
1046    is invoked on the same proc.  This is not what the spec currently
1047    says -- although IMO it arguably should.  (See discussion below.)
1048    
1049    ----------------------------------------------------------------------
1050    Name: Matthias Blume (blume (at) tti - c (dot) org)
1051    Date: 2004/06/17 18:15:00 CDT
1052    Tag: Release_110_46
1053    Description:
1054    
1055    New working version (110.46).  NEW BOOTFILES!
1056    
1057    ----------------------------------------------------------------------
1058    Name: Matthias Blume (blume (at) tti - c (dot) org)
1059    Date: 2004/06/17 17:20:00 CDT
1060    Tag: blume-20040617-timer-unix
1061    Description:
1062    
1063    Changed the interface of structures Timer and Unix to match the most
1064    recent Basis spec.
1065    
1066    In the case of Unix there still seems to be an open/weird issue:
1067    
1068       The {text,bin}{In,Out}streamOf functions are supposed to create
1069       fresh streams whenever they are called -- as opposed to have them
1070       return the same stream every time.  This design is supposed to
1071       prevent space leaks caused by proc values hanging on to streams.
1072    
1073       The reap function, on the other hand, is supposed to close the
1074       streams.  This cannot be done without having a handle on the
1075       stream in proc after all...
1076    
1077    I took the liberty to implement the following stopgap solution:
1078    
1079      The proc value hangs on to the most recently created stream(s).
1080      Reap closes those.  If either or both of the two streams hadn't
1081      been created at all yet, then reap will close the corresponding
1082      file descriptors directly.
1083    
1084    PS: I don't understand the original space leak argument anymore.  If
1085    a proc hangs on to the imperative stream, then I/O operations on those
1086    will advance the state of the cached stream and avoid the space leak.
1087    
1088    ----------------------------------------------------------------------
1089    Name: Matthias Blume (blume (at) tti - c (dot) org)
1090    Date: 2004/05/28 16:45:00 CDT
1091    Tag: blume-20040528-basis
1092    Description:
1093    
1094    Added signature PACK_REAL and exported functor PrimIO.
1095    
1096    ----------------------------------------------------------------------
1097    Name: Matthias Blume (blume (at) tti - c (dot) org)
1098    Date: 2004/05/25 16:00:00 CDT
1099    Tag: blume-20040525-group-owner
1100    Description:
1101    
1102    CM now ignores (but still accepts) the "owner" information in group
1103    descriptions.  The owner of a group is its next enclosing
1104    library. Each group must have a unique owner.  (There is a virtual
1105    "toplevel" library that owns groups which are not nested within a real
1106    library.)  Previously, each group had to explicitly declare its owner,
1107    and CM would check that such a declaration is correct.  The new scheme
1108    is to have CM check that for each group there is precisely one owning
1109    library.
1110    
1111    The advantage of the new scheme is that the programmer no longer needs
1112    to maintain the somewhat annoying owner information.  The downside is
1113    that CM cannot enforce the ownership rule across multiple runs of
1114    CM.make.  Fortunately, enclosing the same group in two different
1115    libraries A and B which are not part of the same program does not
1116    cause real problems.
1117    
1118    ----------------------------------------------------------------------
1119    Name: Matthias Blume (blume (at) tti - c (dot) org)
1120    Date: 2004/05/20 16:00:00 CDT
1121    Tag: blume-20040520-win32
1122    Description:
1123    
1124    Made the win32 version work again.  (Strangely, a misplaced comma had
1125    slipped into win32-process.c which prevented the runtime from being
1126    compiled correctly.)
1127    
1128    Also, included a minor addition to ml-build.bat analogous to what was
1129    done in blume-20040519-ml-build.
1130    
1131    ----------------------------------------------------------------------
1132    Name: Matthias Blume (blume (at) tti - c (dot) org)
1133    Date: 2004/05/19 22:10:00 CDT
1134    Tag: blume-20040519-ml-build
1135    Description:
1136    
1137    Arranged for ml-build to clean up after itself a little bit better.
1138    The script generates a temporary SML source file and compiles it using
1139    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
1140    gets rid of those at the end, so they don't accumulate under .cm.
1141    
1142    This required a minor change to install.sh because the name of the
1143    metadata directory (default: .cm) is actually configurable at
1144    installation time.
1145    
1146    ----------------------------------------------------------------------
1147    Name: Matthias Blume (blume (at) tti - c (dot) org)
1148    Date: 2004/05/18 15:50:00 CDT
1149    Tag: blume-20040518-mkreader
1150    Description:
1151    
1152    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
1153    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
1154    
1155    ----------------------------------------------------------------------
1156    Name: Matthias Blume (blume (at) tti - c (dot) org)
1157    Date: 2004/05/11 14:35:00 CDT
1158    Tag: blume-20040511-win32sock
1159    Description:
1160    
1161    Added previously missing support for many socket-related functions
1162    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
1163    for the voluminous patch!
1164    
1165    (I have not tested this patch under win32 yet.)
1166    
1167    Here is David's e-mail:
1168    
1169    Hi,
1170    
1171    Attached to this email you find a diff against sml/nj 110.45
1172    that will enable socket support under Windows.
1173    
1174    To apply the patch (using unix or cygwin)
1175    1) gunzip runtime.diff.gz
1176    2) "cd" into "src/runtime" in the source tree of a fresh
1177       110.45 installation.
1178    3) patch -p 1 < [your/path/to]runtime.diff
1179    
1180    The code compiles fine but has NOT yet been extensively tested.
1181    I only ran a few tests for basic socket client functionality
1182    (which worked fine).  Especially the functions that use ioctl
1183    are not tested at all and might not work (see below).
1184    
1185    I implemented this since we want to move to a newer version of sml/nj
1186    but need socket support in order to use it.  This is the first time I
1187    even had a look at the sml/nj source,  so please review my changes
1188    before making this part of the distribution!  Here are a few issues
1189    that I think might be better for someone to solve who is more
1190    familiar with the sml/nj source (and socket programming):
1191    
1192    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
1193      exception since I could not figure out what the windows equivalent
1194      of these functions is
1195    
1196    - In sockets-osdep.h there are a some #include statements that are
1197      only used in a few files that include sockets-osdep.h
1198    
1199    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
1200      does not process its return value since I don't know how to
1201      report an error upwards.
1202    
1203    - It would probably be good to have a call to WSACleanup() when
1204      the library is unloaded (if there is such a possibility).
1205      Otherwise I think Windows will take care of this automatically
1206      when the process finishes.
1207    
1208    - I used ioctlsocket() as a replacement for ioctl() but I have
1209      no idea if that is actually the proper replacement on Windows.
1210    
1211    - All these issues are marked in the code by "FIXME" comments.
1212    
1213    We use sml/nj extensively in our products and are quite happy
1214    with it.  I hope this contribution will help you.
1215    
1216    Keep up the good work!
1217    
1218    David
1219    
1220    ----------------------------------------------------------------------
1221    Name: Matthias Blume (blume (at) tti - c (dot) org)
1222    Date: 2004/05/11 14:20:00 CDT
1223    Tag: blume-20040511-installml
1224    Description:
1225    
1226    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
1227    the patch.)
1228    
1229    ----------------------------------------------------------------------
1230    Name: Matthias Blume (blume (at) tti - c (dot) org)
1231    Date: 2004/05/11 14:05:00 CDT
1232    Tag: blume-20040511-nlffi-netbsd
1233    Description:
1234    
1235    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
1236    the patch.)
1237    
1238    ----------------------------------------------------------------------
1239    Name: Matthias Blume (blume (at) tti - c (dot) org)
1240    Date: 2004/05/11 12:05:00 CDT
1241    Tag: blume-20040511-exports
1242    Description:
1243    
1244    As per request by Adam Chlipala <adam@hcoop.net>, extended various
1245    export lists in compiler-related .cm-files.
1246    
1247    ----------------------------------------------------------------------
1248    Name: Matthias Blume (blume (at) tti - c (dot) org)
1249    Date: 2004/05/11 11:35:00 CDT
1250    Tag: blume-20040511-allsource
1251    Description:
1252    
1253    The installer now honors the "src-smlnj" target again, although its meaning
1254    has changed from "all sources required for the compiler" to "all sources
1255    the installer knows about".  In other words, if you enable "src-smlnj"
1256    in the "targets" file, then the installer will pull in sources for
1257    everything.  (Notice that this refers to source code only.  Compiled
1258    code is still only installed for modules that were requested explicitly
1259    or which are required for other modules that were requested explicitly.)
1260    
1261    ----------------------------------------------------------------------
1262    Name: Matthias Blume (blume (at) tti - c (dot) org)
1263    Date: 2004/04/23 17:40:00 CDT
1264    Tag: blume-20040423-ieee-scan
1265    Description:
1266    
1267    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
1268    in the exponent calculation we get INF or ZERO (depending on the mantissa
1269    and the sign of the exponent).
1270    
1271    ----------------------------------------------------------------------
1272    Name: Matthias Blume (blume (at) tti - c (dot) org)
1273    Date: 2004/04/23 10:40:00 CDT
1274    Tag: blume-20040423-ml-build
1275    Description:
1276    
1277    The ml-build script now terminates with a non-0 status when something
1278    goes wrong.
1279    
1280    ----------------------------------------------------------------------
1281    Name: Matthias Blume (blume (at) tti - c (dot) org)
1282    Date: 2004/04/22 16:35:00 CDT
1283    Tag: blume-20040422-Option
1284    Description:
1285    
1286    Made exception Option to be the same as exception Option.Option
1287    (as it should be).
1288    
1289    ----------------------------------------------------------------------
1290    Name: Allen Leung (leunga (at) reservoir (dot) com)
1291    Date: 2004/03/19 14:40:00 EST
1292    Tag: leunga-20040319-cygwin-nlffi
1293    Description:
1294    
1295    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1296    of SML/NJ.  The problem is that
1297    
1298       lib = dlopen(NULL, ...)
1299       f   = dlsym(lib, "malloc");
1300    
1301    does not work on Windows unless we explicitly export symbols
1302    such as 'malloc' during linking.  We fixed this by explicitly
1303    exporting the required symbols with the magic gcc incantation:
1304    
1305        -Wl,--export-all cygwin.def
1306    
1307    where cygwin.def is a file containing all the symbols that
1308    we wish to export.
1309    
1310    I suspect this is a Windows problem and we'll have to
1311    do the same (somehow with windows compilers) when
1312    we build the native win32 version with the system
1313    calls LoadLibrary/GetProcAddress.
1314    
1315    ----------------------------------------------------------------------
1316    Name: Matthias Blume (blume (at) tti - c (dot) org)
1317    Date: 2004/03/04 16:35:00 CST
1318    Tag: blume-20040304-intinf-fmt
1319    Description:
1320    
1321    Fixed problem with IntInf.fmt (sign would show up on the right instead
1322    of on the left for BIN, OCT, and HEX).
1323    
1324    ----------------------------------------------------------------------
1325    Name: Matthias Blume (blume (at) tti - c (dot) org)
1326    Date: 2004/03/04 11:25:00 CST
1327    Tag: blume-20040304-symlinks
1328    Description:
1329    
1330    Fixed problem with installer script (unix only) where bin/ml-yacc and
1331    friends pointed (via symlinks) to absolute locations instead of just
1332    .run-sml.  This was reported by Vesa A Norrman.
1333    
1334    ----------------------------------------------------------------------
1335    Name: Matthias Blume (blume (at) tti - c (dot) org)
1336    Date: 2004/02/13 14:50:00 CST
1337    Tag: Release_110_45
1338    Description:
1339    
1340    New working version (110.45). New bootfiles.
1341    
1342    ----------------------------------------------------------------------
1343    Name: Matthias Blume (blume (at) tti - c (dot) org)
1344    Date: 2004/01/26 15:15:15 CST
1345    Tag: blume-20040126-toplevel
1346    Description:
1347    
1348    Improved handling of exceptions at the interactive toplevel.
1349    
1350    ----------------------------------------------------------------------
1351    Name: Matthias Blume (blume (at) tti - c (dot) org)
1352    Date: 2004/01/26 11:25:00
1353    Tag: blume-20040126-app
1354    Description:
1355    
1356    Type of top-level "app" corrected.
1357    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1358    
1359    ----------------------------------------------------------------------
1360    Name: Matthias Blume (blume (at) tti - c (dot) org)
1361    Date: 2003/11/18 17:10 CST
1362    Tag: blume-20031118-basis-fiddle
1363    Description:
1364    
1365    - changed Timer interface to what might become the spec
1366    - POSIX_FLAGS -> BIT_FLAGS according to spec
1367    - some other minor discrepancies wrt. spec eliminated
1368    
1369    ----------------------------------------------------------------------
1370    Name: Matthias Blume (blume (at) tti - c (dot) org)
1371    Date: 2003/11/06 12:00:00 CST
1372    Tag: Release_110_44
1373    Description:
1374    
1375    New working version (110.44). New bootfiles.
1376    
1377    ----------------------------------------------------------------------
1378    Name: Matthias Blume (blume (at) tti - c (dot) org)
1379    Date: 2003/11/04 11:50:00 CST
1380    Tag: blume-20031104-move-libraries
1381    Description:
1382    
1383    Eliminated the "dont_move_libraries" directive in config/targets.
1384    (The mechanism was broken and could not be fixed easily.  Moreover,
1385    there does not seem to be any reason not to move all libraries into
1386    lib during installation.  I originally implemented this directive as a
1387    backward-compatibility feature when I first introduced the new CM.
1388    Now that things have been stable for a long time and going back to the
1389    old CM is not an option, there is no reason to keep it around.)
1390    
1391    ----------------------------------------------------------------------
1392    Name: Matthias Blume (blume (at) tti - c (dot) org)
1393    Date: 2003/11/03 16:00:00 CST
1394    Tag: blume-20031103-installdir
1395    Description:
1396    
1397    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1398    Richards for pointing out the problem and providing the solution.)
1399    
1400    ----------------------------------------------------------------------
1401    Name: Matthias Blume (blume (at) tti - c (dot) org)
1402    Date: 2003/10/01 17:05:00 CDT
1403    Tag: blume-20031001-lal-mlrisc
1404    Description:
1405    
1406    MLRISC bug fix from Lal.
1407    
1408    ----------------------------------------------------------------------
1409    Name: Matthias Blume (blume (at) tti - c (dot) org)
1410    Date: 2003/09/30 16:10:00 CDT
1411    Tag: blume-20030930-primio-bat
1412    Description:
1413    
1414    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1415    2. Improved .bat files (for Win32 port) to make things work under Win95.
1416       (thanks to Aaron S. Hawley for this one)
1417    
1418    ----------------------------------------------------------------------
1419    Name: Matthias Blume (blume (at) tti - c (dot) org)
1420    Date: 2003/09/26 16:05:00 CDT
1421    Tag: blume-20030926-wrappriv
1422    Description:
1423    
1424    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1425    
1426    ----------------------------------------------------------------------
1427    Name: Matthias Blume (blume (at) tti - c (dot) org)
1428    Date: 2003/09/26 15:00:00 CDT
1429    Tag: blume-20030926-110_43_3
1430    Description:
1431    
1432    - additional cleanup
1433    - version number bump, NEW BOOTFILES
1434    
1435    ----------------------------------------------------------------------
1436    Name: Matthias Blume (blume (at) tti - c (dot) org)
1437    Date: 2003/09/26 12:00:00 CDT
1438    Tag: blume-20030926-ppautoload
1439    Description:
1440    
1441    I modified the read-eval-print loop so that the autoloader gets
1442    invoked whenever the prettyprinter tries to look up a symbol that
1443    is not currently defined in the toplevel environment but which
1444    appears in CM's autoload registry.  As a result, we see far fewer of
1445    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1446    
1447    In addition to this I tried to clean up some pieces of the Basis
1448    implementation (e.g., Socket, Word8Array) in order to prevent other
1449    instances of these ?.Foo.Bar.xxx names from being printed.
1450    
1451    The mechanism that picks names for types still needs some work, though.
1452    (Right now it seems that if there is a type A.t which is defined to
1453    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1454    "?.B.u" although the perhaps more sensible solution would be to use
1455    "A.t" in this case.  In other words, the prettyprinter should follow
1456    a chain of DEFtycs not farther than there are corresponding toplevel
1457    names in the current environment.)
1458    
1459    ----------------------------------------------------------------------
1460    Name: Matthias Blume (blume (at) tti - c (dot) org)
1461    Date: 2003/09/24 16:31:00 CDT
1462    Tag: blume-20030924-installer
1463    Description:
1464    
1465    Another installer tweak:  All the ML code for the installer is now
1466    compiled during CMB.make and put into a little library called
1467    $smlnj/installer.cm.  The installation then simply invokes
1468    
1469       sml -m $smlnj/installer.cm
1470    
1471    and everything happens automagically.
1472    
1473    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1474    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1475            and BINDIR.
1476    
1477    The new scheme guarantees that the ML code responsible for the installation
1478    is in sync with the APIs of the main system.  Also, the installer is
1479    somewhat faster because the installer script is precompiled.
1480    
1481    ----------------------------------------------------------------------
1482    Name: Matthias Blume (blume (at) tti - c (dot) org)
1483    Date: 2003/09/24 15:35:00 CDT
1484    Tag: blume-20030924-synsock
1485    Description:
1486    
1487    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1488    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1489    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1490    implementations of non-blocking operations from CML's Socket
1491    structure.
1492    
1493    ----------------------------------------------------------------------
1494    Name: Matthias Blume (blume (at) tti - c (dot) org)
1495    Date: 2003/09/24 15:10:05 CDT
1496    Tag: blume-20030924-sockets
1497    Description:
1498    
1499    1. Fixed SOCKET API and implementation to match Basis spec.
1500       This required changing the internal representation of sockets to one
1501       that remembers (for each socket file descriptor) whether it is currently
1502       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1503       call is made only if the state actually needs to change).
1504    
1505    2. OS-specific details of sockets were moved into separate files, thus
1506       making it possible to unify the bulk of the socket implementations
1507       between Unix and Win32.
1508    
1509    3. CML's socket API changed accordingly.
1510       (Note that we need to remove non-blocking functions from this API
1511       since they are redundant in the case of CML!)
1512    
1513    4. CML's socket implementation now makes use of non-blocking functions
1514       provided by Basis, thus removing all OS-dependent code from this part
1515       of CML.
1516    
1517    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1518    
1519    ----------------------------------------------------------------------
1520    Name: Matthias Blume (blume (at) tti - c (dot) org)
1521    Date: 2003/09/22 12:10:00 CDT
1522    Tag: blume-20030922-110_43_2
1523    Description:
1524    
1525    Made a new interim version and bootfiles for developer's bootstrapping
1526    convenience.
1527    
1528       110.43.2 -- NEW BOOTFILES
1529    
1530    ----------------------------------------------------------------------
1531    Name: Matthias Blume (blume (at) tti - c (dot) org)
1532    Date: 2003/09/19 15:55:00 CDT
1533    Tag: blume-20030919-cmdir
1534    Description:
1535    
1536    1. new-install.sh -> install.sh
1537    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1538    3. tweaked installer so that another name instead of .cm can be chosen
1539       at install time (by setting the CM_DIR_ARC environment variable
1540       during installation); once installation is complete, the name is
1541       fixed
1542    
1543    ----------------------------------------------------------------------
1544    Name: Matthias Blume (blume (at) tti - c (dot) org)
1545    Date: 2003/09/18 16:00:00 CDT
1546    Tag: blume-20030918-110_43_1
1547    Description:
1548    
1549    Made a new interim version and bootfiles for developer's bootstrapping
1550    convenience.
1551    
1552       110.43.1 -- NEW BOOTFILES
1553    
1554    ----------------------------------------------------------------------
1555    Name: Matthias Blume (blume (at) tti - c (dot) org)
1556    Date: 2003/09/18 15:20:00 CDT
1557    Tag: blume-20030918-misc
1558    Description:
1559    
1560    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1561       this seems to be controversial at the moment) and used those in
1562       Posix.ProcEnv.times.
1563    
1564    2. Added Time.{from,to}Nanoseconds to Time.
1565    
1566    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1567       For example, fromLargeInt never needs to look at more than 3 "big
1568       digits" to get its 53 bits of precision.
1569    
1570    ----------------------------------------------------------------------
1571    Name: Matthias Blume (blume (at) tti - c (dot) org)
1572    Date: 2003/09/17 16:30:00 CDT
1573    Tag: blume-20030917-real32-slices
1574    Description:
1575    
1576    Added an entry to the primitive environment
1577    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1578    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1579    
1580    Removed some of the "magic" constants in real64.sml and replaced them
1581    with code that generates these values from their corresponding
1582    integer counterparts.
1583    
1584    Made all(?) the slice-related changes to the Basis and made everything
1585    compile again...
1586    
1587    ----------------------------------------------------------------------
1588    Name: Matthias Blume (blume (at) tti - c (dot) org)
1589    Date: 2003/09/15 17:45:00 CDT
1590    Tag: blume-20030915-rbase
1591    Description:
1592    
1593    Fixed bug in Real.fromLargeInt.
1594    
1595    ----------------------------------------------------------------------
1596    Name: Matthias Blume (blume (at) tti - c (dot) org)
1597    Date: 2003/09/13 18:11:00 CDT
1598    Tag: blume-20030913-libinstall
1599    Description:
1600    
1601    Minor bugfix in config/libinstall (set anchor with path to
1602    standalone tool after installing it, otherwise libraries that
1603    need ml-lex or ml-yacc won't compile the first time the installer
1604    runs).
1605    
1606    ----------------------------------------------------------------------
1607    Name: Matthias Blume (blume (at) tti - c (dot) org)
1608    Date: 2003/09/12 11:45:00 CDT
1609    Tag: blume-20030912-various
1610    Description:
1611    
1612     - fixed bug in Real.toLargeInt
1613     - fixed bug in Posix.ProcEnv.times
1614     - changed inputLine functions to return an option
1615     - minor installer improvements / bugfixes
1616     - changed default @SMLalloc parameter for x86/celeron to 64k
1617    
1618    ----------------------------------------------------------------------
1619    Name: Matthias Blume (blume (at) tti - c (dot) org)
1620    Date: 2003/09/09 22:00:00 CDT
1621    Tag: Release_110_43
1622    Description:
1623    
1624        New working release 110.43.  New bootfiles.
1625    
1626    ----------------------------------------------------------------------
1627    Name: Matthias Blume (blume (at) tti - c (dot) org)
1628    Date: 2003/09/09 19:20:00 CDT
1629    Tag: blume-20030909-installer
1630    Description:
1631    
1632    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1633    Modified config/install.bat to take advantage of it.  Also modified
1634    config/install.sh (and called it config/new-install.sh) to take advantage
1635    of it on Unix systems.  (The SML code is (supposed to be) platform-
1636    independent.)
1637    
1638    The installer can now install everything under Win32
1639    as well as under *nix as long as it compiles.
1640    
1641    Other changes:
1642    
1643      - made CML compile again under Win32
1644      - made eXene compile under Win32 (by providing a fake structure UnixSock
1645        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1646      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
1647        same as type int;  under Win32 it isn't
1648      - fixed some slice-related problems in the win32-specific parts of CML
1649      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
1650        (under Win32, the volume name is case-insensitive, and the
1651        OS.Path code compares volume names for equality)
1652    
1653    ----------------------------------------------------------------------
1654    Name: Matthias Blume (blume (at) tti - c (dot) org)
1655    Date: 2003/09/08 11:55:00 CDT
1656    Tag: blume-20030908-fullpath
1657    Description:
1658    
1659    Made Win32 version of OS.FileSys.fullPath return current directory
1660    when given an empty string.  This is what the spec says, and incidentally,
1661    CM depends on it.  (CM otherwise goes into an infinite loop in certain
1662    cases when presented with the name of a non-existing .cm file.)
1663    
1664    ----------------------------------------------------------------------
1665    Name: Matthias Blume (blume (at) tti - c (dot) org)
1666    Date: 2003/09/04 16:30:00 CDT
1667    Tag: blume-20030905-slices-etc
1668    Description:
1669    
1670      1. Changed interface to vectors and arrays in Basis to match
1671         (draft) Basis spec.
1672      2. Added signatures and implementations of slices according to
1673         Basis spec.
1674      3. Edited source code throughout the system to make it compile again
1675         under 1. and 2.  (In some cases code had to be added to have it
1676         match the new signatures.)
1677      4. MLRISC should be backward-compatible: the copies of the originals
1678         of files that needed to change under 3. were retained, the .cm files
1679         check the compiler version number and use old versions when
1680         appropriate.
1681      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1682         dirstream -> string option (in accordance with Basis spec).
1683      6. When generating code that counts lines, ml-lex used function
1684         CharVector.foldli, taking advantage of its old interface.
1685         This has been replaced with the corresponding code from
1686         CharVectorSlice.  (html-lex must be re-lexed!)
1687      7. BitArray in smlnj-lib/Util has been extended/modified to match the
1688         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
1689      8. Removed temporary additions (fromInternal, toInternal) from the
1690         (now obsolete) IntInf in smlnj-lib/Util.
1691      9. Cleaned up structure Byte.
1692     10. Added localOffset, scan, and fromString to Date (according to spec).
1693         Cleaned/corrected implementation of Date.
1694         (Still need to check for correctness; implement better canonicalizeDate.)
1695     11. Added "scan" to signature IEEE_REAL.
1696     12. Some improvements to IntInf [in particular: efficiency-hack for
1697         mod and rem when second operand is 2 (for parity checks).]
1698     13. Changed representation of type Time.time, using a single IntInf.int
1699         value counting microseconds.  This considerably simplified the
1700         implementation of structure Time.  We now support negative time
1701         values; scan and fromString handle signs.
1702     14. Functor PrimIO now takes two additional arguments (VectorSlice and
1703         ArraySlice).
1704    
1705    ----------------------------------------------------------------------
1706    Name: Matthias Blume (blume (at) tti - c (dot) org)
1707    Date: 2003/08/28 17:00:00 CDT
1708    Tag: blume-20030828-intinf
1709    Description:
1710    
1711    This is a major update which comes with a version number bump
1712    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1713    and an implementation of IntInf in the Basis.
1714    
1715    There are a fairly large number of related changes and updates throughout
1716    the system:
1717    
1718    Basis:
1719            - Implemented IntInf.
1720            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1721            - Added some missing Real64 operations, most notably Real.toLargeInt.
1722            - Added FixedInt as a synonym for Int32.
1723    
1724    compiler:
1725            * Added support for a built-in intinf type.
1726              - literals
1727              - pattern matching
1728              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1729              - overloading on literals and operations
1730    
1731            This required adding a primitive type intinf, some additional
1732            primops, and implementations for several non-trivial intinf
1733            operations in Core.  (The intinf type is completely abstract
1734            to the compiler; all operations get delegated back to the Core.)
1735    
1736            * Intinf equality is handled by polyequal.  However, the compiler
1737            does not print its usual warning in this case (since polyequal
1738            is the right thing to do there).
1739    
1740            * Improved the organization of structure InlineT.
1741    
1742            * A word about conversion primops:
1743            If conversions involving intinf do not cancel out during
1744            CPS contract, then the compiler must insert calls to Core functions.
1745            Since all core access must be resolved already during the FLINT
1746            translate phase, it would be too late a the time of CPS contract
1747            to add new Core calls.  For this reason, conversion primops
1748            for intinf carry two arguments:  1. the numeric argument that
1749            they are supposed to convert, and 2. the Core function that
1750            can help with this conversion if necessary.  If CPS contract
1751            eliminates a primop, then the associated Core function becomes
1752            dead and goes away.  Intinf conversion primops that do not get
1753            eliminated by CPS contract get rewritten into calls of their
1754            core functions by a separate, new phase.
1755    
1756    interactive system:
1757            - Control.Print.intinfDepth controls max length of intinf constants
1758            being printed. (Analogous to Control.Print.stringDepth.)
1759            - Cleanup in printutil and pputil: got rid of unused stuff and
1760            duplicates; replaced some of the code with code that makes better
1761            use of library functionality.
1762    
1763    CM:
1764            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1765            are no longer silent.
1766    
1767    CKIT:
1768            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1769            in favor of LargeInt -- which is now the same as IntInf.
1770            CKIT-knowledgable people should check whether this is what's
1771            intended and otherwise change things back to using Int32 or
1772            FixedInt.
1773    
1774    Throughout the code:
1775            Started using IntInf.int literals and built-in operations
1776            (e.g., comparison with 0) where this seems appropriate.
1777    
1778    
1779    ----------------------------------------------------------------------
1780    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1781    Date: 2003/08/13 11:36:00 CDT
1782    Tag: dbm-20030813-mcz-merge1
1783    Description:
1784    
1785    Merging changes from the mcz-branch development branch into trunk.
1786    These changes involve replacement of the emulated old prettyprinter
1787    interface with direct use of the SML/NJ Lib PP library, and fixing
1788    of a couple of bugs (895, 1186) relating to error messages.  A new
1789    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1790    has been added.
1791    
1792    ----------------------------------------------------------------------
1793    Name: Matthias Blume (blume (at) tti - c (dot) org)
1794    Date: 2003/08/11 15:45:00 CDT
1795    Tag: blume-20030811-windows
1796    Description:
1797    
1798    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1799    
1800        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1801    
1802    This patch restores SML/NJ's ability to run under win32.  There are a
1803    number of changes, including fixes for several bugs that had gone
1804    unnoticed until now:
1805    
1806      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1807        trying to run the win32 version from within cygwin.)
1808    
1809      - There are a number of simple .bat scripts that substitute for their
1810        corresponding Unix shell-scripts.  (See below.)
1811    
1812      - The internals of ml-build have been modified slightly.  The main
1813        difference is that instead of calling ".link-sml" (or link-sml.bat)
1814        using OS.Process.system, the ML process delegates this task back
1815        to the script.  Otherwise problems arise in mixed environments such
1816        as Cygwin where scripts look and work like Unix scripts, but
1817        where OS.Process.system cannot run them.
1818    
1819      - In CM, the srcpath pickler used native pathname syntax -- which
1820        is incorrect in the case of cross-compilation.  The new pickle format
1821        is independent of platform-specific naming conventions.
1822    
1823      - Path configuration files (such as lib/pathconfig) can now choose
1824        between native and standard syntax.  Placing a line of the form
1825    
1826            standard!
1827    
1828        into the file causes all subsequent paths to be interpreted using
1829        CM standard pathname syntax (= Unix conventions); a line
1830    
1831            native!
1832    
1833        switches back to native style.  This was needed so that
1834        path config files can be written portably, see src/system/pathconfig.
1835    
1836      - Runtime system:
1837    
1838          - win32-filesys.c:  get_file_time and set_file_time now
1839            access modification time, not creation time.
1840    
1841          - I/O code made aware of new array representation.
1842    
1843          - Bug fixes in X86.prim.masm.
1844    
1845       - src/system/makeml made aware of win32. (For use under cygwin
1846            and other Unix-environments for windows.)
1847    
1848       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1849         which caused BinIO.inputAll to fail consistently.
1850    
1851    .bat scripts:
1852    
1853       Windows .bat scripts assume that SMLNJ_HOME is defined.
1854    
1855       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1856         applications (sml, ml-yacc, ml-lex).
1857       - ml-build.bat: analogous to ml-build.
1858       - config\install.bat:  Analogous to config/install.sh.  This requires
1859         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1860         (nmake etc. must be on the path, and vcvars32 must have been run.)
1861         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1862         and the bootfile hierarchy must have been unpacked under
1863         sml.boot.x86-win32.
1864         The script is very primitive and does a poor job at error checking.
1865         It only installs the base system, ml-lex, and ml-yacc.  No other
1866         libraries are being installed (i.e., you get only those that
1867         are part of the compiler.)
1868       - link-sml.bat:  analogous to .link-sml, but not currently used
1869    
1870    Unrelated bug fixes:
1871    
1872      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1873        types.
1874      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1875        otherwise.)
1876    
1877    ----------------------------------------------------------------------
1878    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1879    Date: 2003/06/17
1880    Tag: macqueen-20030617-bug895
1881    Description:
1882    
1883    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1884    Tag will be used for new development branch (mcz-branch) for
1885    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1886    
1887    ----------------------------------------------------------------------
1888    Name: Matthias Blume (blume (at) tti - c (dot) org)
1889    Date: 2003/05/27 16:55:00 CDT
1890    Tag: blume-20030527-polyeq
1891    Description:
1892    
1893    Tried to eliminated most cases of polymorphic equality.
1894    
1895    ----------------------------------------------------------------------
1896    Name: Matthias Blume (blume (at) tti - c (dot) org)
1897    Date: 2003/05/21 17:45:00 CDT
1898    Tag: blume-20030517-complete
1899    Description:
1900    
1901    Two changes:
1902    
1903      1. Added a flag for controlling whether non-exhaustive bindings will
1904         be treated as errors (default is false).
1905      2. Cleaned up the *entire* source tree so that CMB.make goes through
1906         without a single non-exhaustive match- or bind warning.
1907    
1908    ----------------------------------------------------------------------
1909    Name: Matthias Blume (blume (at) tti - c (dot) org)
1910    Date: 2003/05/17 10:20:00 CDT
1911    Tag: blume-20030517-absyn
1912    Description:
1913    
1914    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
1915    
1916       This mainly affects the quality of error messages.  However, some
1917       of the code is now more straightforward than before. (Treatment of
1918       the above four constructs in translate.sml is much simpler than
1919       the "macro-expansion" that was going on before.  Plus, the mach-
1920       compiler no longer gets invoked just to be able to compile an
1921       if-expression.)
1922    
1923    2. The ErrorMsg.Error exception is now caught and absorbed by the
1924       interactive loop.
1925    
1926    ----------------------------------------------------------------------
1927    Name: Allen Leung
1928    Date: 2003/05/16 13:05:00 CDT
1929    Tag: leunga-20030516-cygwin-runtime
1930    Description:
1931    
1932        Ported the runtime system to cygwin, which uses the unix
1933    x86-unix bin files.  Missing/buggy features:
1934    
1935       o getnetbyname, getnetbyaddr: these functions seem to be missing in
1936         the Cygwin library.
1937       o Ctrl-C handling may be flaky.
1938       o Windows system calls and Windows I/O are not supported.
1939    
1940       A new set of binfiles is located at:
1941    
1942         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
1943    
1944       This is only needed for bootstrapping the cygwin version of smlnj.
1945       Other x86 versions can use the existing binfiles.
1946    
1947    ----------------------------------------------------------------------
1948    Name: Matthias Blume
1949    Date: 2003/04/08 15:42:00 CDT
1950    Tag: blume-20030408-listpair
1951    Description:
1952    
1953    1. Added a target 'mlrisc' to installer.
1954    
1955    2. Added missing elements to structure ListPair.
1956    
1957    ----------------------------------------------------------------------
1958    Name: Allen Leung
1959    Date: 2003/01/07 10:40:00 EST
1960    Tag: leunga-20030107-int-rem
1961    Description:
1962    
1963    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
1964    The arguments to the SUBL instruction were swapped.
1965    
1966    ----------------------------------------------------------------------
1967    Name: Matthias Blume
1968    Date: 2002/12/12 16:25:00 EST
1969    Tag: blume-20021212-risc-ra
1970    Description:
1971    
1972    Fixed a serious bug in the rewrite code for FP spilling/reloading that
1973    sent the RA into an infinite loop when floating point registers get
1974    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
1975    110.37 and 110.38.)
1976    There was another set of potential problems related to the handling of
1977    MLRISC annotations (but those did not yet cause real problems, apparently).
1978    
1979    ----------------------------------------------------------------------
1980    Name: Matthias Blume
1981    Date: 2002/12/06 22:40:00 EST
1982    Tag: blume-20021206-cm-fileid
1983    Description:
1984    
1985    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
1986    This fixes the problem of CM getting confused by files that suddenly
1987    change their identity (e.g., by getting unlinked and recreated by some
1988    text editor such as vi).  There might be a better/cheaper/cleaner way
1989    of doing this, but for now this will have to do.
1990    
1991    ----------------------------------------------------------------------
1992    Name: Matthias Blume
1993    Date: 2002/10/28 09:50:00 EST
1994    Tag: blume-20021028-typecheck
1995    Description:
1996    
1997    Exported structure Typecheck from $smlnj/viscomp/core.cm.
1998    
1999    ----------------------------------------------------------------------
2000    Name: Matthias Blume
2001    Date: 2002/10/17 09:10:00 EDT
2002    Tag: Release_110_42
2003    Description:
2004    
2005    In good old tradition, there has been a slight hiccup so that we have
2006    to patch 110.42 after the fact.   The old release tag has been replaced
2007    (see below).
2008    
2009    The change solves a problem with two competing approaches the
2010    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
2011    each other's way.
2012    
2013    This change only affects the runtime system code and the installer script.
2014    (No new bootfiles.)
2015    
2016    ----------------------------------------------------------------------
2017    Name: Matthias Blume
2018    Date: 2002/10/16 12:00:00 EDT
2019    Tag: Release_110_42_removed
2020    Description:
2021    
2022    New working release.  New bootfiles.
2023    
2024    ----------------------------------------------------------------------
2025    Name: Matthias Blume
2026    Date: 2002/10/10 13:10:00 EDT
2027    Tag: blume-20021010-ppc-divs
2028    Description:
2029    
2030    The mltree operator DIVS must be implemented with an overflow check on
2031    the PPC because the hardware indicates divide-by-zero using "overflow" as
2032    well.
2033    
2034    ----------------------------------------------------------------------
2035    Name: Matthias Blume
2036    Date: 2002/07/23 11:45:00
2037    Tag: blume-20020723-smlnj-home
2038    Description:
2039    
2040    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
2041    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
2042    CM_PATHCONFIG is also set), the path configuration file is assumed
2043    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
2044    the entire tree to some other place and everything will "just work".
2045    
2046    (Companion commands such as ml-build and ml-makedepend also sense this
2047    variable.)
2048    
2049    ----------------------------------------------------------------------
2050    Name: Matthias Blume
2051    Date: 2002/07/12 21:19:00 EDT
2052    Tag: blume-20020712-liveness
2053    Description:
2054    
2055    Exported two useful "step" functions from liveness module (MLRISC).
2056    
2057    ----------------------------------------------------------------------
2058    Name: Matthias Blume
2059    Date: 2002/07/05 16:00 EDT
2060    Tag: Release_110_41
2061    Description:
2062    
2063    New working release.  New bootfiles.
2064    
2065    ----------------------------------------------------------------------
2066    Name: Matthias Blume
2067    Date: 2002/07/05 10:25:00 EDT
2068    Tag: blume-20020705-btimp
2069    Description:
2070    
2071    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
2072    other clients can set up backtracing support.
2073    
2074    ----------------------------------------------------------------------
2075    Name: Matthias Blume
2076    Date: 2002/06/25 17:23:00 EDT
2077    Tag: blume-20020625-fpmax
2078    Description:
2079    
2080    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
2081    case.  (The sense of the isNaN test was reversed -- which made min and
2082    max always return their first argument.)
2083    
2084    ----------------------------------------------------------------------
2085    Name: Matthias Blume
2086    Date: 2002/06/11
2087    Tag: blume-20020611-unixpath
2088    Description:
2089    
2090    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
2091    
2092    ----------------------------------------------------------------------
2093    Name: Matthias Blume
2094    Date: 2002/06/10 16:35:00 EDT
2095    Tag: blume-20020610-ieeereal
2096    Description:
2097    
2098    I back-ported my implementation of IEEEReal.fromString from the
2099    idlbasis-devel branch so that we can test it.
2100    
2101    Another small change is that ppDec tries to give more information
2102    than just "<sig>" in the case of functors.  However, this code is
2103    broken in some mysterious way if the functor's body's signature
2104    has not been declared by ascription but gets inferred from the
2105    implementation.  This needs fixing...
2106    
2107    ----------------------------------------------------------------------
2108    Name: Matthias Blume
2109    Date: 2002/05/31
2110    Tag: blume-20020531-btrace-mode
2111    Description:
2112    
2113    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
2114    the wayside when I switched over to using Controls everywhere.)
2115    
2116    ----------------------------------------------------------------------
2117    Name: Lal George
2118    Date: 2002/05/23 12:21:40 EDT
2119    Tag: george-20020523-visual-labels
2120    Description:
2121    
2122    Labels are now displayed in the graphical output to make
2123    the fall-through and target blocks obvious.
2124    
2125    ----------------------------------------------------------------------
2126    Name: Matthias Blume
2127    Date: 2002/05/22 11:03:00 EDT
2128    Tag: blume-20020522-shrink
2129    Description:
2130    
2131    John tweaked yesterday's fix for 1131 to handle an out-of-memory
2132    situation that comes up when allocating huge arrays.
2133    
2134    ----------------------------------------------------------------------
2135    Name: Matthias Blume
2136    Date: 2002/05/21 16:00:00 EDT
2137    Tag: Release_110_40
2138    Description:
2139    
2140    New working release (110.40).  New bootfiles.
2141    
2142    [Also: John Reppy fixed GC bug 1131.]
2143    
2144    ----------------------------------------------------------------------
2145    Name: Matthias Blume
2146    Date: 2002/05/21 12:35:00 EDT
2147    Tag: blume-20020521-cmdoc
2148    Description:
2149    
2150    CM documentation update.
2151    
2152    ----------------------------------------------------------------------
2153    Name: Matthias Blume
2154    Date: 2002/05/21 10:55:00 EDT
2155    Tag: blume-20020521-misc
2156    Description:
2157    
2158    - John tweaked runtime to be silent on heap export (except when
2159      GC messages are on).
2160    - I added a few more things (cross-compiling versions of CMB) to
2161      config/preloads (as suggestions).
2162    
2163    ----------------------------------------------------------------------
2164    Name: Matthias Blume
2165    Date: 2002/05/20 22:25:00 EDT
2166    Tag: blume-20020520-controls
2167    Description:
2168    
2169    - Added ControlUtil structure to control-lib.cm.
2170    - Use it throughout.
2171    - Used Controls facility to define MLRISC controls (as opposed to
2172      registering MLRISC control ref cells with Controls after the fact)
2173    - Fixed messed-up controls priorities.
2174    
2175    * Removed again all the stuff from config/preloads that one wouldn't
2176      be able to preload at the time the initial heap image is built.
2177      (Many libraries, e.g., CML, do not exist yet at this time.  The
2178       only libraries that can be preloaded via config/preloads are those
2179       that come bundled with the bootfiles.)
2180    
2181    ----------------------------------------------------------------------
2182    Name: Matthias Blume
2183    Date: 2002/05/20 10:59:00 EDT
2184    Tag: blume-20020520-preloads
2185    Description:
2186    
2187    Added a lot of commented-out suggestions for things to be included
2188    in config/preloads.
2189    
2190    ----------------------------------------------------------------------
2191    Name: Allen Leung
2192    Date: 2002/05/18 14:20:00 EDT
2193    Tag: leunga-20020518-mdl
2194    Description:
2195    
2196    o Made the mdl tool stuff compile and run again.
2197    o I've disabled all the stuff that depends on RTL specifications; they
2198      are all badly broken anyway.
2199    
2200    ----------------------------------------------------------------------
2201    Name: Matthias Blume
2202    Date: 2002/05/17 16:49:00 EDT
2203    Tag: blume-20020517-controls
2204    Description:
2205    
2206    0. John Reppy made several modifications to the SML/NJ library.
2207       In particular, there is a shiny new controls-lib.cm.
2208    
2209    1. Pushed new controls interface through compiler so that everything
2210       compiles again.
2211    
2212    2. Added FormatComb and FORMAT_COMB to the CML version of the
2213       SML/NJ library (so that CML compiles again).
2214    
2215    3. Modified init scripts because XXX_DEFAULT environment variables
2216       are no longer with us.  (Boot-time initialization is now done
2217       using the same environment variables that are also used for
2218       startup-time initialization of controls.)
2219    
2220    ----------------------------------------------------------------------
2221    Name: Lal George
2222    Date: 2002/05/15 09:20:10 EDT
2223    Tag: george-20020515-pseudo-op-decls
2224    Description:
2225    
2226            All pseudo-ops emitted before the first segment declaration
2227    such as TEXT, DATA, and BSS directives are assumed to be global
2228    declarations and are emitted first in the assembly file. This is
2229    useful in a number of situations where one has pseudo-ops that are not
2230    specific to any segment, and also works around the constraint that one
2231    cannot have client pseudo-ops in the TEXT segment.
2232    
2233            Because no segment is associated with these declarations it is
2234    an error to allocate any space or objects before the first segment
2235    directive and an exception will be raised. However, we cannot make
2236    this check for client pseudo-ops.
2237    
2238            These top level declarations are a field in the CFG graph_info.
2239    In theory you can continue to add to this field after the CFG has been
2240    built -- provided you know what you are doing;-)
2241    
2242    ----------------------------------------------------------------------
2243    Name: Matthias Blume
2244    Date: 2002/05/13 16:40:00 EDT
2245    Tag: blume-20020513-pp-etc
2246    Description:
2247    
2248    A few minor bugfixes:
2249    
2250      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
2251        (Bogus printouts for redefined bindings still occur.  Compiler
2252        bug should no longer occur now.  We need to redo the prettyprinter
2253        from scratch.)
2254    
2255      - CM pathname printer now also adds escape sequences for ( and )
2256    
2257      - commend and docu fixes for ml-nlffi
2258    
2259    ----------------------------------------------------------------------
2260    Name: Matthias Blume
2261    Date: 2002/05/10 16:40:00 EDT
2262    Tag: blume-20020510-erg-textio
2263    Description:
2264    
2265    Applied the following bugfix provided by Emden Gansner:
2266    
2267        Output is corrupted when outputSubstr is used rather than output.
2268    
2269        The problem occurs when a substring
2270    
2271            ss = (s, dataStart, dataLen)
2272    
2273        where dataStart > 0, fills a stream buffer with avail bytes left.
2274        avail bytes of s, starting at index dataStart, are copied into the
2275        buffer, the buffer is flushed, and then the remaining dataLen-avail
2276        bytes of ss are copied into the beginning of the buffer. Instead of
2277        starting this copy at index dataStart+avail in s, the current code
2278        starts the copy at index avail.
2279    
2280        Fix:
2281        In text-io-fn.sml, change line 695 from
2282             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2283        to
2284             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2285    
2286    ----------------------------------------------------------------------
2287    Name: Matthias Blume
2288    Date: 2002/04/12 13:55:00 EDT
2289    Tag: blume-20020412-assyntax
2290    Description:
2291    
2292    1. Grabbed newer assyntax.h from the XFree86 project.
2293    2. Fiddled with how to compile X86.prim.asm without warnings.
2294    3. (Very) Minor cleanup in CM.
2295    
2296    ----------------------------------------------------------------------
2297    Name: Matthias Blume
2298    Date: 2002/04/01 (no joke!) 17:07:00 EST
2299    Tag: blume-20020401-x86div
2300    Description:
2301    
2302    Added full support for div/mod/rem/quot on the x86, using the machine
2303    instruction's two results (without clumsily recomputing the remainder)
2304    directly where appropriate.
2305    
2306    Some more extensive power-of-two support was added to the x86 instruction
2307    selector (avoiding expensive divs, mods, and muls where they can be
2308    replaced with cheaper shifts and masks).  However, this sort of thing
2309    ought to be done earlier, e.g., within the CPS optimizer so that
2310    all architectures benefit from it.
2311    
2312    The compiler compiles to a fixed point, but changes might be somewhat
2313    fragile nevertheless.  Please, report any strange things that you might
2314    see wrt. div/mod/quot/rem...
2315    
2316    ----------------------------------------------------------------------
2317    Name: Matthias Blume
2318    Date: 2002/03/29 17:22:00
2319    Tag: blume-20020329-div
2320    Description:
2321    
2322    Fixed my broken div/mod logic.  Unfortunately, this means that the
2323    inline code for div/mod now has one more comparison than before.
2324    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2325    The problem was with quotient = 0, because that alone does not tell
2326    us which way the rounding went.  One then has to look at whether
2327    remainder and divisor have the same sign...  :(
2328    
2329    Anyway, I replaced the bootfiles with fresh ones...
2330    
2331    ----------------------------------------------------------------------
2332    Name: Matthias Blume
2333    Date: 2002/03/29 14:10:00 EST
2334    Tag: blume-20020329-inlprims
2335    Description:
2336    
2337    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2338    
2339    Primops have changed. This means that the bin/boot-file formats have
2340    changed as well.
2341    
2342    To make sure that there is no confusion, I made a new version.
2343    
2344    
2345    CHANGES:
2346    
2347    * removed REMT from mltree (remainder should never overflow).
2348    
2349    * added primops to deal with divisions of all flavors to the frontend
2350    
2351    * handled these primops all the way through so they map to their respective
2352      MLRISC support
2353    
2354    * used these primops in the implementation of Int, Int32, Word, Word32
2355    
2356    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2357    
2358    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2359    
2360    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2361      floating point
2362    
2363    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2364      and Real (but Real.abs maps to a separate floating-point-only primop)
2365    
2366    
2367    TODO items:
2368    
2369    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2370      instructions because the machine instruction encoder cannot handle
2371      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2372      REM in the first place, and REM is supported by the x86 machine coder.)
2373    
2374    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2375      selection phase.  (The two can be streamlined because the hardware
2376      delivers both quotient and remainder at the same time anyway.)
2377    
2378    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2379      (Currently the behavior is inconsistent both across architectures and
2380      wrt. the draft Basis spec.)
2381    
2382    * Word8 should eventually be handled natively, too.
2383    
2384    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2385      as if there currently is no execution path that could trigger it in
2386      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2387      hold for things like multiplication and division.)
2388    
2389    ----------------------------------------------------------------------
2390    Name: Matthias Blume
2391    Date: 2002/03/27 16:27:00 EST
2392    Tag: blume-20020327-mlrisc-divisions
2393    Description:
2394    
2395    Added support for all four division operations (ML's div, mod, quot,
2396    and rem) to MLRISC.  In the course of doing so, I also rationalized
2397    the naming (no more annoying switch-around of DIV and QUOT), by
2398    parameterizing the operation by div_rounding_mode (which can be either
2399    DIV_TO_ZERO or DIV_TO_NEGINF).
2400    
2401    The generic MLTreeGen functor takes care of compiling all four
2402    operations down to only round-to-zero div.
2403    
2404    Missing pieces:
2405    
2406      * Doing something smarter than relying on MLTreeGen on architectures
2407        like, e.g., the x86 where hardware division delivers both quotient and
2408        remainder at the same time.  With this, the implementation of the
2409        round-to-neginf operations could be further streamlined.
2410    
2411      * Remove inlining support for div/mod/rem from the frontend and replace it
2412        with primops that get carried through to the backend.  Do this for all
2413        int and word types.
2414    
2415    ----------------------------------------------------------------------
2416    Name: Matthias Blume
2417    Date: 2002/03/25 17:25:00 EST
2418    Tag: blume-20020325-divmod
2419    Description:
2420    
2421    I improved (hopefully without breaking them) the implementation of Int.div,
2422    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2423    advantage of the following observations:
2424    
2425      Let  q = x quot y      r = x rem y
2426           d = x div  y      m = x mod y
2427    
2428    where "quot" is the round-to-zero version of integer division that
2429    hardware usually provides.  Then we have:
2430    
2431         r = x - q * y        where neither the * nor the - will overflow
2432         d = if q >= 0 orelse x = q * y then q else q - 1
2433                              where neither the * nor the - will overflow
2434         m = if q >= 0 orelse r = 0 then r else r + y
2435                              where the + will not overflow
2436    
2437    This results in substantial simplification of the generated code.
2438    The following table shows the number of CFG nodes and edges generated
2439    for
2440            fun f (x, y) = x OPER y
2441            (* with OPER \in div, mod, quot, rem *)
2442    
2443    
2444        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2445        --------------------------------------------------------
2446         div |         24 |         39 |         12 |         16
2447         mod |         41 |         71 |         12 |         16
2448        quot |          8 |         10 |          8 |         10
2449         rem |         10 |         14 |          8 |         10
2450    
2451    
2452    ----------------------------------------------------------------------
2453    Name: Matthias Blume
2454    Date: 2002/03/25 22:06:00 EST
2455    Tag: blume-20020325-cprotobug
2456    Description:
2457    
2458    Fixed a bug in cproto (c prototype decoder).
2459    
2460    ----------------------------------------------------------------------
2461    Name: Matthias Blume
2462    Date: 2002/03/25 16:00:00 EST
2463    Tag: blume-20020325-raw-primops
2464    Description:
2465    
2466    I did some cleanup to Allen's new primop code and
2467    replaced yesterday's bootfiles with new ones.
2468    (But they are stored in the same place.)
2469    
2470    ----------------------------------------------------------------------
2471    Name: Matthias Blume
2472    Date: 2002/03/24 22:40:00 EST
2473    Tag: blume-20020324-bootfiles
2474    Description:
2475    
2476    Made the bootfiles that Allen asked for.
2477    
2478    ----------------------------------------------------------------------
2479    Name: Allen Leung
2480    Date: 2002/03/23 15:50:00 EST
2481    Tag: leunga-20020323-flint-cps-rcc-primops
2482    Description:
2483    
2484      1. Changes to FLINT primops:
2485    
2486        (* make a call to a C-function;
2487         * The primop carries C function prototype information and specifies
2488         * which of its (ML-) arguments are floating point. C prototype
2489         * information is for use by the backend, ML information is for
2490         * use by the CPS converter. *)
2491      | RAW_CCALL of { c_proto: CTypes.c_proto,
2492                       ml_args: ccall_type list,
2493                       ml_res_opt: ccall_type option,
2494                       reentrant : bool
2495                     } option
2496       (* Allocate uninitialized storage on the heap.
2497        * The record is meant to hold short-lived C objects, i.e., they
2498        * are not ML pointers.  With the tag, the representation is
2499        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2500        *)
2501      | RAW_RECORD of {tag:bool,sz:int}
2502      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2503    
2504      2.  These CPS primops are now overloaded:
2505    
2506           rawload of {kind:numkind}
2507           rawstore of {kind:numkind}
2508    
2509          The one argument form is:
2510    
2511             rawload {kind} address
2512    
2513          The two argument form is:
2514    
2515             rawload {kind} [ml object, byte-offset]
2516    
2517      3. RAW_CCALL/RCC now takes two extra arguments:
2518    
2519         a. The first is whether the C call is reentrant, i.e., whether
2520            ML state should be saved and restored.
2521         b. The second argument is a string argument specifying the name of
2522            library and the C function.
2523    
2524         These things are currently not handled in the code generator, yet.
2525    
2526      4. In CProto,
2527    
2528         An encoding type of "bool" means "ml object" and is mapped into
2529         C prototype of PTR.  Note that "bool" is different than "string",
2530         even though "string" is also mapped into PTR, because "bool"
2531         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2532    
2533      5. Pickler/unpicker
2534    
2535         Changed to handle RAW_RECORD and newest RAW_CCALL
2536    
2537      6. MLRiscGen,
2538    
2539         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2540         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2541            in the file CodeGen/cpscompile/cps-c-calls.sml
2542    
2543      7. Added the conditional move operator
2544    
2545             condmove of branch
2546    
2547         to cps.  Generation of this is still buggy so it is currently
2548         disabled.
2549    
2550    ----------------------------------------------------------------------
2551    Name: Lal George
2552    Date: 2002/03/22 14:18:25 EST
2553    Tag: george-20020322-cps-branch-prob
2554    Description:
2555    
2556    Implemented the Ball-Larus branch prediction-heuristics, and
2557    incorporated graphical viewers for control flow graphs.
2558    
2559    Ball-Larus Heuristics:
2560    ---------------------
2561    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2562    
2563    By design it uses the Dempster-Shafer theory for combining
2564    probabilities.  For example, in the function:
2565    
2566        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2567    
2568    the ball-larus heuristics predicts that the n=0 is unlikely
2569    (OH-heuristic), and the 'then' branch is unlikely because of the
2570    RH-heuristic -- giving the 'then' branch an even lower combined
2571    probability using the Dempster-Shafer theory.
2572    
2573    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2574    probability of the 'then' branch because of the loop in the else
2575    branch.
2576    
2577    
2578    Graphical Viewing:
2579    ------------------
2580    I merely plugged in Allen's graphical viewers into the compiler. The
2581    additional code is not much. At the top level, saying:
2582    
2583            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2584    
2585    will display the graphical view of the control flow graph just before
2586    back-patching.  daVinci must be in your path for this to work. If
2587    daVinci is not available, then the default viewer can be changed
2588    using:
2589    
2590            Control.MLRISC.getString "viewer"
2591    
2592    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2593    course, these viewers must be in your path.
2594    
2595    The above will display the compilation unit at the level of clusters,
2596    many of which are small, boring, and un-interesting. Also setting:
2597    
2598            Control.MLRISC.getInt "cfg-graphical-view_size"
2599    
2600    will display clusters that are larger than the value set by the above.
2601    
2602    
2603    ----------------------------------------------------------------------
2604    Name: Matthias Blume
2605    Date: 2002/03/21 22:20:00 EST
2606    Tag: blume-20020321-kmp-bugfix
2607    Description:
2608    
2609    Changed the interface to the KMP routine in PreString and fixed
2610    a minor bug in one place where it was used.
2611    
2612    ----------------------------------------------------------------------
2613    Name: Allen Leung
2614    Date: 2002/03/21 20:30:00 EST
2615    Tag: leunga-20020321-cfg
2616    Description:
2617    
2618      Fixed a potential problem in cfg edge splitting.
2619    
2620    ----------------------------------------------------------------------
2621    Name: Allen Leung
2622    Date: 2002/03/21 17:15:00 EST
2623    Tag: leunga-20020321-x86-fp-cfg
2624    Description:
2625    
2626      1. Recoded the buggy parts of x86-fp.
2627    
2628         a. All the block reordering code has been removed.
2629            We now depend on the block placement phases to do this work.
2630    
2631         b. Critical edge splitting code has been simplified and moved into the
2632            CFG modules, as where they belong.
2633    
2634         Both of these were quite buggy and complex.  The code is now much, much
2635         simpler.
2636    
2637      2. X86 backend.
2638    
2639         a. Added instructions for 64-bit support.  Instruction selection for
2640            64-bit has not been committed, however, since that
2641            requires changes to MLTREE which haven't been approved by
2642            Lal and John.
2643    
2644         b. Added support for FUCOMI and FUCOMIP when generating code for
2645            PentiumPro and above.  We only generate these instructions in
2646            the fast-fp mode.
2647    
2648         c. Added cases for JP and JNP in X86FreqProps.
2649    
2650      3. CFG
2651    
2652         CFG now has a bunch of methods for edge splitting and merging.
2653    
2654      4. Machine description.
2655    
2656         John's simplification of MLTREE_BASIS.fcond broke a few machine
2657         description things:
2658    
2659         rtl-build.{sig,sml} and hppa.mdl fixed.
2660    
2661         NOTE: the machine description stuff in the repository is still broken.
2662               Again, I can't put my fixes in because that involves
2663               changes to MLTREE.
2664    
2665    ----------------------------------------------------------------------
2666    Name: Matthias Blume
2667    Date: 2002/03/20 15:55:00 EST
2668    Tag: blume-20020320-kmp
2669    Description:
2670    
2671    Implemented Knuth-Morris-Pratt string matching in PreString and used
2672    it for String.isSubstring, Substring.isSubstring, and
2673    Substring.position.
2674    
2675    (Might need some stress-testing.  Simple examples worked fine.)
2676    
2677    ----------------------------------------------------------------------
2678    Name: Matthias Blume
2679    Date: 2002/03/19 16:37:00 EST
2680    Tag: blume-20020319-witnesses
2681    Description:
2682    
2683    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2684    
2685    This implements a generic mechanism for changing constness qualifiers
2686    anywhere within big C types without resorting to outright "casts".
2687    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2688    modify the constness at the outermost level.)
2689    The implementation of "convert" is based on the idea of "witness"
2690    values -- values that are not used by the operation but whose types
2691    "testify" to their applicability.  On the implementation side, "convert"
2692    is simply a projection (returning its second curried argument).  With
2693    cross-module inlining, it should not result in any machine code being
2694    generated.
2695    
2696    ----------------------------------------------------------------------
2697    Name: Matthias Blume
2698    Date: 2002/03/15 16:40:00 EST
2699    Tag: blume-20020315-basis
2700    Description:
2701    
2702    Provided (preliminary?) implementations for
2703    
2704      {String,Substring}.{concatWith,isSuffix,isSubstring}
2705    
2706    and
2707    
2708      Substring.full
2709    
2710    Those are in the Basis spec but they were missing in SML/NJ.
2711    
2712    ----------------------------------------------------------------------
2713    Name: Matthias Blume
2714    Date: 2002/03/14 21:30:00 EST
2715    Tag: blume-20020314-controls
2716    Description:
2717    
2718    Controls:
2719    ---------
2720    
2721    1. Factored out the recently-added Controls : CONTROLS stuff and put
2722       it into its own library $/controls-lib.cm.  The source tree for
2723       this is under src/smlnj-lib/Controls.
2724    
2725    2. Changed the names of types and functions in this interface, so they
2726       make a bit more "sense":
2727    
2728          module -> registry
2729          'a registry -> 'a group
2730    
2731    3. The interface now deals in ref cells only.  The getter/setter interface
2732       is (mostly) gone.
2733    
2734    4. Added a function that lets one register an already-existing ref cell.
2735    
2736    5. Made the corresponding modifications to the rest of the code so that
2737       everything compiles again.
2738    
2739    6. Changed the implementation of Controls.MLRISC back to something closer
2740       to the original.  In particular, this module (and therefore MLRISC)
2741       does not depend on Controls.  There now is some link-time code in
2742       int-sys.sml that registers the MLRISC controls with the Controls
2743       module.
2744    
2745    CM:
2746    ---
2747    
2748      * One can now specify the lambda-split aggressiveness in init.cmi.
2749    
2750    ----------------------------------------------------------------------
2751    Name: Allen Leung
2752    Date: 2002/03/13 17:30:00 EST
2753    Tag: leunga-20020313-x86-fp-unary
2754    Description:
2755    
2756    Bug fix for:
2757    
2758    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2759    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2760    > - fun f(x,(y,z)) = Real.~ y;
2761    > [autoloading]
2762    > [autoloading done]
2763    >       fchsl   (%eax), 184(%esp)
2764    > Error: MLRisc bug: X86MCEmitter.emitInstr
2765    >
2766    > uncaught exception Error
2767    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2768    
2769    The problem was that the code generator did not generate any fp registers
2770    in this case, and the ra didn't know that it needed to run the X86FP phase to
2771    translate the pseudo fp instruction.   This only happened with unary fp
2772    operators in certain situations.
2773    
2774    ----------------------------------------------------------------------
2775    Name: Matthias Blume
2776    Date: 2002/03/13 14:00:00 EST
2777    Tag: blume-20020313-overload-etc
2778    Description:
2779    
2780    1. Added _overload as a synonym for overload for backward compatibility.
2781       (Control.overloadKW must be true for either version to be accepted.)
2782    
2783    2. Fixed bug in install script that caused more things to be installed
2784       than what was requested in config/targets.
2785    
2786    3. Made CM aware of the (_)overload construct so that autoloading
2787       works.
2788    
2789    ----------------------------------------------------------------------
2790    Name: Matthias Blume
2791    Date: 2002/03/12 22:03:00 EST
2792    Tag: blume-20020312-url
2793    Description:
2794    
2795    Forgot to update BOOT and srcarchiveurl.
2796    
2797    ----------------------------------------------------------------------
2798    Name: Matthias Blume
2799    Date: 2002/03/12 17:30:00 EST
2800    Tag: blume-20020312-version110392
2801    Description:
2802    
2803    Yet another version number bump (because of small changes to the
2804    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2805    
2806    Changes:
2807    
2808      The new pid generation scheme described a few weeks ago was overly
2809      complicated.  I implemented a new mechanism that is simpler and
2810      provides a bit more "stability":  Once CM has seen a compilation
2811      unit, it keeps its identity constant (as long as you do not delete
2812      those crucial CM/GUID/* files).  This means that when you change
2813      an interface, compile, then go back to the old interface, and
2814      compile again, you arrive at the original pid.
2815    
2816      There now also is a mechanism that instructs CM to use the plain
2817      environment hash as a module's pid (effectively making its GUID
2818      the empty string).  For this, "noguid" must be specified as an
2819      option to the .sml file in question within its .cm file.
2820      This is most useful for code that is being generated by tools such
2821      as ml-nlffigen (because during development programmers tend to
2822      erase the tool's entire output directory tree including CM's cached
2823      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2824      revert to the old, broken behavior of SML/NJ, but in specific cases
2825      where there is no danger of interface confusion, its use is ok
2826      (I think).
2827    
2828      ml-nlffigen by default generates "noguid" annotations.  They can be
2829      turned off by specifying -guid in its command line.
2830    
2831    ----------------------------------------------------------------------
2832    Name: Lal George
2833    Date: 2002/03/12 12 14:42:36 EST
2834    Tag: george-20020312-frequency-computation
2835    Description:
2836    
2837    Integrated jump chaining and static block frequency into the
2838    compiler. More details and numbers later.
2839    
2840    ----------------------------------------------------------------------
2841    Name: Lal George
2842    Date: 2002/03/11 11 22:38:53 EST
2843    Tag: george-20020311-jump-chain-elim
2844    Description:
2845    
2846    Tested the jump chain elimination on all architectures (except the
2847    hppa).  This is on by default right now and is profitable for the
2848    alpha and x86, however, it may not be profitable for the sparc and ppc
2849    when compiling the compiler.
2850    
2851    The gc test will typically jump to a label at the end of the cluster,
2852    where there is another jump to an external cluster containing the actual
2853    code to invoke gc. This is to allow factoring of common gc invocation
2854    sequences. That is to say, we generate:
2855    
2856            f:
2857               testgc
2858               ja   L1      % jump if above to L1
2859    
2860            L1:
2861               jmp L2
2862    
2863    
2864    After jump chain elimination the 'ja L1' instructions is converted to
2865    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2866    up being implemented in their long form (if L2 is far away) using:
2867    
2868            jbe     L3      % jump if below or equal to L3
2869            jmp     L2
2870         L3:
2871            ...
2872    
2873    
2874    For large compilation units L2  may be far away.
2875    
2876    
2877    ----------------------------------------------------------------------
2878    Name: Matthias Blume
2879    Date: 2002/03/11 13:30:00 EST
2880    Tag: blume-20020311-mltreeeval
2881    Description:
2882    
2883    A functor parameter was missing.
2884    
2885    ----------------------------------------------------------------------
2886    Name: Allen Leung
2887    Date: 2002/03/11 10:30:00 EST
2888    Tag: leunga-20020311-runtime-string0
2889    Description:
2890    
2891       The representation of the empty string now points to a
2892    legal null terminated C string instead of unit.  It is now possible
2893    to convert an ML string into C string with InlineT.CharVector.getData.
2894    This compiles into one single machine instruction.
2895    
2896    ----------------------------------------------------------------------
2897    Name: Allen Leung
2898    Date: 2002/03/10 23:55:00 EST
2899    Tag: leunga-20020310-x86-call
2900    Description:
2901    
2902       Added machine generation for CALL instruction (relative displacement mode)
2903    
2904    ----------------------------------------------------------------------
2905    Name: Matthias Blume
2906    Date: 2002/03/08 16:05:00
2907    Tag: blume-20020308-entrypoints
2908    Description:
2909    
2910    Version number bumped to 110.39.1.  NEW BOOTFILES!
2911    
2912    Entrypoints: non-zero offset into a code object where execution should begin.
2913    
2914    - Added the notion of an entrypoint to CodeObj.
2915    - Added reading/writing of entrypoint info to Binfile.
2916    - Made runtime system bootloader aware of entrypoints.
2917    - Use the address of the label of the first function given to mlriscGen
2918      as the entrypoint.  This address is currently always 0, but it will
2919      not be 0 once we turn on block placement.
2920    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
2921      with entry points) from mlriscGen.
2922    
2923    ----------------------------------------------------------------------
2924    Name: Allen Leung
2925    Date: 2002/03/07 20:45:00 EST
2926    Tag: leunga-20020307-x86-cmov
2927    Description:
2928    
2929       Bug fixes for CMOVcc on x86.
2930    
2931       1. Added machine code generation for CMOVcc
2932       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
2933       3. CMOVcc cannot have an immediate operand as argument.
2934    
2935    ----------------------------------------------------------------------
2936    Name: Matthias Blume
2937    Date: 2002/03/07 16:15:00 EST
2938    Tag: blume-20020307-controls
2939    Description:
2940    
2941    This is a very large but mostly boring patch which makes (almost)
2942    every tuneable compiler knob (i.e., pretty much everything under
2943    Control.* plus a few other things) configurable via both the command
2944    line and environment variables in the style CM did its configuration
2945    until now.
2946    
2947    Try starting sml with '-h' (or, if you are brave, '-H')
2948    
2949    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
2950    implements the underlying generic mechanism.
2951    
2952    The interface to some of the existing such facilities has changed somewhat.
2953    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
2954    (The getFoo interface is still there for backward-compatibility, but its
2955    use is deprecated.)
2956    
2957    The ml-build script passes -Cxxx=yyy command-line arguments through so
2958    that one can now twiddle the compiler settings when using this "batch"
2959    compiler.
2960    
2961    TODO items:
2962    
2963    We should go through and throw out all controls that are no longer
2964    connected to anything.  Moreover, we should go through and provide
2965    meaningful (and correct!) documentation strings for those controls
2966    that still are connected.
2967    
2968    Currently, multiple calls to Controls.new are accepted (only the first
2969    has any effect).  Eventually we should make sure that every control
2970    is being made (via Controls.new) exactly once.  Future access can then
2971    be done using Controls.acc.
2972    
2973    Finally, it would probably be a good idea to use the getter-setter
2974    interface to controls rather than ref cells.  For the time being, both
2975    styles are provided by the Controls module, but getter-setter pairs are
2976    better if thread-safety is of any concern because they can be wrapped.
2977    
2978    *****************************************
2979    
2980    One bug fix: The function blockPlacement in three of the MLRISC
2981    backpatch files used to be hard-wired to one of two possibilities at
2982    link time (according to the value of the placementFlag).  But (I
2983    think) it should rather sense the flag every time.
2984    
2985    *****************************************
2986    
2987    Other assorted changes (by other people who did not supply a HISTORY entry):
2988    
2989    1. the cross-module inliner now works much better (Monnier)
2990    2. representation of weights, frequencies, and probabilities in MLRISC
2991       changed in preparation of using those for weighted block placement
2992       (Reppy, George)
2993    
2994    ----------------------------------------------------------------------
2995    Name: Lal George
2996    Date: 2002/03/07 14:44:24 EST 2002
2997    Tag: george-20020307-weighted-block-placement
2998    
2999    Tested the weighted block placement optimization on all architectures
3000    (except the hppa) using AMPL to generate the block and edge frequencies.
3001    Changes were required in the machine properties to correctly
3002    categorize trap instructions. There is an MLRISC flag
3003    "weighted-block-placement" that can be used to enable weighted block
3004    placement, but this will be ineffective without block/edge
3005    frequencies (coming soon).
3006    
3007    
3008    ----------------------------------------------------------------------
3009    Name: Lal George
3010    Date: 2002/03/05 17:24:48 EST
3011    Tag: george-20020305-linkage-cluster
3012    
3013    In order to support the block placement optimization, a new cluster
3014    is generated as the very first cluster (called the linkage cluster).
3015    It contains a single jump to the 'real' entry point for the compilation
3016    unit. Block placement has no effect on the linkage cluster itself, but
3017    all the other clusters  have full freedom in the manner in which they
3018    reorder blocks or functions.
3019    
3020    On the x86 the typical linkage code that is generated is:
3021       ----------------------
3022            .align 2
3023       L0:
3024            addl    $L1-L0, 72(%esp)
3025            jmp     L1
3026    
3027    
3028            .align  2
3029       L1:
3030       ----------------------
3031    
3032    72(%esp) is the memory location for the stdlink register. This
3033    must contain the address of the CPS function being called. In the
3034    above example, it contains the address of  L0; before
3035    calling L1 (the real entry point for the compilation unit), it
3036    must contain the address for L1, and hence
3037    
3038            addl $L1-L0, 72(%esp)
3039    
3040    I have tested this on all architectures except the hppa.The increase
3041    in code size is of course negligible
3042    
3043    ----------------------------------------------------------------------
3044    Name: Allen Leung
3045    Date: 2002/03/03 13:20:00 EST
3046    Tag: leunga-20020303-mlrisc-tools
3047    
3048      Added #[ ... ] expressions to mlrisc tools
3049    
3050    ----------------------------------------------------------------------
3051    Name: Matthias Blume
3052    Date: 2002/02/27 12:29:00 EST
3053    Tag: blume-20020227-cdebug
3054    Description:
3055    
3056    - made types in structure C and C_Debug to be equal
3057    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
3058    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
3059    
3060    ----------------------------------------------------------------------
3061    Name: Matthias Blume
3062    Date: 2002/02/26 12:00:00 EST
3063    Tag: blume-20020226-ffi
3064    Description:
3065    
3066    1. Fixed a minor bug in CM's "noweb" tool:
3067       If numbering is turned off, then truly don't number (i.e., do not
3068       supply the -L option to noweb).  The previous behavior was to supply
3069       -L'' -- which caused noweb to use the "default" line numbering scheme.
3070       Thanks to Chris Richards for pointing this out (and supplying the fix).
3071    
3072    2. Once again, I reworked some aspects of the FFI:
3073    
3074       A. The incomplete/complete type business:
3075    
3076       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
3077         gone!
3078       - ML types representing an incomplete type are now *equal* to
3079         ML types representing their corresponding complete types (just like
3080         in C).  This is still safe because ml-nlffigen will not generate
3081         RTTI for incomplete types, nor will it generate functions that
3082         require access to such RTTI.   But when ML code generated from both
3083         incomplete and complete versions of the C type meet, the ML types
3084         are trivially interoperable.
3085    
3086         NOTE:  These changes restore the full generality of the translation
3087         (which was previously lost when I eliminated functorization)!
3088    
3089       B. Enum types:
3090    
3091       - Structure C now has a type constructor "enum" that is similar to
3092         how the "su" constructor works.  However, "enum" is not a phantom
3093         type because each "T enum" has values (and is isomorphic to
3094         MLRep.Signed.int).
3095       - There are generic access operations for enum objects (using
3096         MLRep.Signed.int).
3097       - ml-nlffigen will generate a structure E_foo for each "enum foo".
3098         * The structure contains the definition of type "mlrep" (the ML-side
3099         representation type of the enum).  Normally, mlrep is the same
3100         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
3101         then mlrep will be defined as a datatype -- thus facilitating
3102         pattern matching on mlrep values.
3103         ("-ec" will be suppressed if there are duplicate values in an
3104          enumeration.)
3105         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
3106         will be generated for each C enum constant xxx.
3107         * Conversion functions m2i and i2m convert between mlrep and
3108         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
3109         * Coversion functions c and ml convert between mlrep and "tag enum".
3110         * Access functions (get/set) fetch and store mlrep values.
3111       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
3112         enumerations are merged into one single enumeration represented by
3113         structure E_'.
3114    
3115    ----------------------------------------------------------------------
3116    Name: Allen Leung
3117    Date: 2002/02/25 04:45:00 EST
3118    Tag: leunga-20020225-cps-spill
3119    
3120    This is a new implementation of the CPS spill phase.
3121    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
3122    In case of problems, replace it with the old file spill.sml
3123    
3124    The current compiler runs into some serious performance problems when
3125    constructing a large record.  This can happen when we try to compile a
3126    structure with many items.  Even a very simple structure like the following
3127    makes the compiler slow down.
3128    
3129        structure Foo = struct
3130           val x_1 = 0w1 : Word32.int
3131           val x_2 = 0w2 : Word32.int
3132           val x_3 = 0w3 : Word32.int
3133           ...
3134           val x_N = 0wN : Word32.int
3135        end
3136    
3137    The following table shows the compile time, from N=1000 to N=4000,
3138    with the old compiler:
3139    
3140    N
3141    1000   CPS 100 spill                           0.04u  0.00s  0.00g
3142           MLRISC ra                               0.06u  0.00s  0.05g
3143              (spills = 0 reloads = 0)
3144           TOTAL                                   0.63u  0.07s  0.21g
3145    
3146    1100   CPS 100 spill                           8.25u  0.32s  0.64g
3147           MLRISC ra                               5.68u  0.59s  3.93g
3148              (spills = 0 reloads = 0)
3149           TOTAL                                   14.71u  0.99s  4.81g
3150    
3151    1500   CPS 100 spill                           58.55u  2.34s  1.74g
3152           MLRISC ra                               5.54u  0.65s  3.91g
3153              (spills = 543 reloads = 1082)
3154           TOTAL                                   65.40u  3.13s  6.00g
3155    
3156    2000   CPS 100 spill                           126.69u  4.84s  3.08g
3157           MLRISC ra                               0.80u  0.10s  0.55g
3158              (spills = 42 reloads = 84)
3159           TOTAL                                   129.42u  5.10s  4.13g
3160    
3161    3000   CPS 100 spill                           675.59u  19.03s  11.64g
3162           MLRISC ra                               2.69u  0.27s  1.38g
3163              (spills = 62 reloads = 124)
3164           TOTAL                                   682.48u  19.61s  13.99g
3165    
3166    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
3167           MLRISC ra                               4.96u  0.27s  2.72g
3168              (spills = 85 reloads = 170)
3169           TOTAL                                   2375.26u  57.21s  48.00g
3170    
3171    As you can see the old cps spill module suffers from some serious
3172    performance problem.  But since I cannot decipher the old code fully,
3173    instead of patching the problems up, I'm reimplementing it
3174    with a different algorithm.  The new code is more modular,
3175    smaller when compiled, and substantially faster
3176    (O(n log n) time and O(n) space).  Timing of the new spill module:
3177    
3178    4000  CPS 100 spill                           0.02u  0.00s  0.00g
3179          MLRISC ra                               0.25u  0.02s  0.15g
3180             (spills=1 reloads=3)
3181          TOTAL                                   7.74u  0.34s  1.62g
3182    
3183    Implementation details:
3184    
3185    As far as I can tell, the purpose of the CPS spill module is to make sure the
3186    number of live variables at any program point (the bandwidth)
3187    does not exceed a certain limit, which is determined by the
3188    size of the spill area.
3189    
3190    When the bandwidth is too large, we decrease the register pressure by
3191    packing live variables into spill records.  How we achieve this is
3192    completely different than what we did in the old code.
3193    
3194    First, there is something about the MLRiscGen code generator
3195    that we should be aware of:
3196    
3197    o MLRiscGen performs code motion!
3198    
3199       In particular, it will move floating point computations and
3200       address computations involving only the heap pointer to
3201       their use sites (if there is only a single use).
3202       What this means is that if we have a CPS record construction
3203       statement
3204    
3205           RECORD(k,vl,w,e)
3206    
3207       we should never count the new record address w as live if w
3208       has only one use (which is often the case).
3209    
3210       We should do something similar to floating point, but the transformation
3211       there is much more complex, so I won't deal with that.
3212    
3213    Secondly, there are now two new cps primops at our disposal:
3214    
3215     1. rawrecord of record_kind option
3216        This pure operator allocates some uninitialized storage from the heap.
3217        There are two forms:
3218    
3219         rawrecord NONE [INT n]  allocates a tagless record of length n
3220         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
3221                                     and initializes the tag.
3222    
3223     2. rawupdate of cty
3224          rawupdate cty (v,i,x)
3225          Assigns to x to the ith component of record v.
3226          The storelist is not updated.
3227    
3228    We use these new primops for both spilling and increment record construction.
3229    
3230     1. Spilling.
3231    
3232        This is implemented with a linear scan algorithm (but generalized
3233        to trees).  The algorithm will create a single spill record at the
3234        beginning of the cps function and use rawupdate to spill to it,
3235        and SELECT or SELp to reload from it.  So both spills and reloads
3236        are fine-grain operations.  In contrast, in the old algorithm
3237        "spills" have to be bundled together in records.
3238    
3239        Ideally, we should sink the spill record construction to where
3240        it is needed.  We can even split the spill record into multiple ones
3241        at the places where they are needed.  But CPS is not a good
3242        representation for global code motion, so I'll keep it simple and
3243        am not attempting this.
3244    
3245     2. Incremental record construction (aka record splitting).
3246    
3247        Long records with many component values which are simulatenously live
3248        (recall that single use record addresses are not considered to
3249         be live) are constructed with rawrecord and rawupdate.
3250        We allocate space on the heap with rawrecord first, then gradually
3251        fill it in with rawupdate.  This is the technique suggested to me
3252        by Matthias.
3253    
3254        Some restrictions on when this is applicable:
3255        1. It is not a VECTOR record.  The code generator currently does not handle
3256           this case. VECTOR record uses double indirection like arrays.
3257        2. All the record component values are defined in the same "basic block"
3258           as the record constructor.  This is to prevent speculative
3259           record construction.
3260    
3261    ----------------------------------------------------------------------
3262    Name: Allen Leung
3263    Date: 2002/02/22 01:02:00 EST
3264    Tag: leunga-20020222-mlrisc-tools
3265    
3266    Minor bug fixes in the parser and rewriter
3267    
3268    ----------------------------------------------------------------------
3269    Name: Allen Leung
3270    Date: 2002/02/21 20:20:00 EST
3271    Tag: leunga-20020221-peephole
3272    
3273    Regenerated the peephole files.  Some contained typos in the specification
3274    and some didn't compile because of pretty printing bugs in the old version
3275    of 'nowhere'.
3276    
3277    ----------------------------------------------------------------------
3278    Name: Allen Leung
3279    Date: 2002/02/19 20:20:00 EST
3280    Tag: leunga-20020219-mlrisc-tools
3281    Description:
3282    
3283       Minor bug fixes to the mlrisc-tools library:
3284    
3285       1.  Fixed up parsing colon suffixed keywords
3286       2.  Added the ability to shut the error messages up
3287       3.  Reimplemented the pretty printer and fixed up/improved
3288           the pretty printing of handle and -> types.
3289       4.  Fixed up generation of literal symbols in the nowhere tool.
3290       5.  Added some SML keywords to to sml.sty
3291    
3292    ----------------------------------------------------------------------
3293    Name: Matthias Blume
3294    Date: 2002/02/19 16:20:00 EST
3295    Tag: blume-20020219-cmffi
3296    Description:
3297    
3298    A wild mix of changes, some minor, some major:
3299    
3300    * All C FFI-related libraries are now anchored under $c:
3301        $/c.cm      --> $c/c.cm
3302        $/c-int.cm  --> $c/internals/c-int.cm
3303        $/memory.cm --> $c/memory/memory.cm
3304    
3305    * "make" tool (in CM) now treats its argument pathname slightly
3306      differently:
3307        1. If the native expansion is an absolute name, then before invoking
3308           the "make" command on it, CM will apply OS.Path.mkRelative
3309           (with relativeTo = OS.FileSys.getDir()) to it.
3310        2. The argument will be passed through to subsequent phases of CM
3311           processing without "going native".  In particular, if the argument
3312           was an anchored path, then "make" will not lose track of that anchor.
3313    
3314    * Compiler backends now "know" their respective C calling conventions
3315      instead of having to be told about it by ml-nlffigen.  This relieves
3316      ml-nlffigen from one of its burdens.
3317    
3318    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3319    
3320    * Export C_DEBUG and C_Debug from $c/c.cm.
3321    
3322    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3323      subtyping relationship between incomplete pointers and their complete
3324      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3325      with the convention of instantiating 'o with ('t, 'c) obj whenever
3326      the pointer target type is complete.  In the incomplete case, 'o
3327      will be instantiated with some "'c iobj" -- a type obtained by
3328      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3329    
3330      Operations that work on both incomplete and complete pointer types are
3331      typed as taking an 'o ptr while operations that require the target to
3332      be known are typed as taking some ('t, 'c) obj ptr.
3333    
3334      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3335      where void is an eqtype without any values.  This makes it possible
3336      to work on voidptr values using functions meant to operate on light
3337      incomplete pointers.
3338    
3339    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3340      been vastly simplified.
3341    
3342    ----------------------------------------------------------------------
3343    Name: Matthias Blume
3344    Date: 2002/02/19 10:48:00 EST
3345    Tag: blume-20020219-pqfix
3346    Description:
3347    
3348    Applied Chris Okasaki's bug fix for priority queues.
3349    
3350    ----------------------------------------------------------------------
3351    Name: Matthias Blume
3352    Date: 2002/02/15 17:05:00
3353    Tag: Release_110_39
3354    Description:
3355    
3356    Last-minute retagging is becoming a tradition... :-(
3357    
3358    This is the working release 110.39.
3359    
3360    ----------------------------------------------------------------------
3361    Name: Matthias Blume
3362    Date: 2002/02/15 16:00:00 EST
3363    Tag: Release_110_39-orig
3364    Description:
3365    
3366    Working release 110.39.  New bootfiles.
3367    
3368    (Update: There was a small bug in the installer so it wouldn't work
3369    with all shells.  So I retagged. -Matthias)
3370    
3371    ----------------------------------------------------------------------
3372    Name: Matthias Blume
3373    Date: 2002/02/15 14:17:00 EST
3374    Tag: blume-20020215-showbindings
3375    Description:
3376    
3377    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3378    the latter can be useful for exploring what bindings are available at
3379    the interactive prompt.  (The first function returns only the list
3380    of symbols that are really bound, the second prints those but also the
3381    ones that CM's autoloading mechanism knows about.)
3382    
3383    ----------------------------------------------------------------------
3384    Name: Matthias Blume
3385    Date: 2002/02/15 12:08:00 EST
3386    Tag: blume-20020215-iptrs
3387    Description:
3388    
3389    Two improvements to ml-nlffigen:
3390    
3391      1. Write files only if they do not exist or if their current contents
3392         do not coincide with what's being written.  (That is, avoid messing
3393         with the time stamps unless absolutely necessary.)
3394    
3395      2. Implement a "repository" mechanism for generated files related
3396         to "incomplete pointer types".   See the README file for details.
3397    
3398    ----------------------------------------------------------------------
3399    Name: Matthias Blume
3400    Date: 2002/02/14 11:50:00 EST
3401    Tag: blume-20020214-quote
3402    Description:
3403    
3404    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3405    because of the new and improved tag generation scheme.  (Thanks to Allen
3406    Leung for pointing it out.)
3407    
3408    ----------------------------------------------------------------------
3409    Name: Lal George
3410    Date: 2002/02/14 09:55:27 EST 2002
3411    Tag: george-20020214-isabelle-bug
3412    Description:
3413    
3414    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3415    of Isabelle on the x86.
3416    
3417    From Allen:
3418    -----------
3419     I've found the problem:
3420    
3421         in ra-core.sml, I use the counter "blocked" to keep track of the
3422         true number of elements in the freeze queue.  When the counter goes
3423         to zero, I skip examining the queue.  But I've messed up the
3424         bookkeeping in combine():
3425    
3426             else ();
3427             case !ucol of
3428               PSEUDO => (if !cntv > 0 then
3429                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3430                                        ^^^^^^^^^^^^^^^^^^^^^^^
3431                      moveu := mergeMoveList(!movev, !moveu)
3432                     )
3433                  else ();
3434    
3435         combine() is called to coalesce two nodes u and v.
3436         I think I was thinking that if the move counts of u and v are both
3437         greater than zero then after they are coalesced then one node is
3438         removed from the freeze queue.  Apparently I was thinking that
3439         both u and v are of low degree, but that's clearly not necessarily true.
3440    
3441    
3442    02/12/2002:
3443        Here's the patch.  HOL now compiles.
3444    
3445        I don't know how this impact on performance (compile
3446        time or runtime).  This bug caused the RA (especially on the x86)
3447        to go thru the potential spill phase when there are still nodes on the
3448        freeze queue.
3449    
3450    
3451    
3452    
3453    ----------------------------------------------------------------------
3454    Name: Matthias Blume
3455    Date: 2002/02/13 22:40:00 EST
3456    Tag: blume-20020213-fptr-rtti
3457    Description:
3458    
3459    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3460    updates.
3461    
3462    ----------------------------------------------------------------------
3463    Name: Matthias Blume
3464    Date: 2002/02/13 16:41:00 EST
3465    Tag: blume-20020213-cmlpq
3466    Description:
3467    
3468    Added new priority queue export symbols (which have just been added to
3469    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3470    not compile and the installer would choke.)
3471    
3472    ----------------------------------------------------------------------
3473    Name: Matthias Blume
3474    Date: 2002/02/13 16:15:00 EST
3475    Tag: blume-20020213-various
3476    Description:
3477    
3478    1. More tweaks to ml-nlffigen:
3479    
3480       - better internal datastructures (resulting in slight speedup)
3481       - "-match" option requires exact match
3482       - "localized" gensym counters (untagged structs/unions nested within
3483         other structs/unions or within typedefs get a fresh counter; their
3484         tag will be prefixed by a concatenation of their parents' tags)
3485       - bug fixes (related to calculation of transitive closure of types
3486         to be included in the output)
3487    
3488    2. Minor Basis updates:
3489    
3490       - added implementations for List.collate and Option.app
3491    
3492    ----------------------------------------------------------------------
3493    Name: Matthias Blume
3494    Date: 2002/02/11 15:55:00 EST
3495    Tag: blume-20020211-gensym
3496    Description:
3497    
3498    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3499    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3500    names (ML structure names that correspond to unnamed C structs, unions,
3501    and enums), so that separate runs of ml-nlffigen do not clash.
3502    
3503    ----------------------------------------------------------------------
3504    Name: Matthias Blume
3505    Date: 2002/02/11 12:05:00 EST
3506    Tag: blume-20020211-gensml
3507    Description:
3508    
3509    A quick fix for a problem with GenSML (in the pgraph-util library):
3510    Make generation of toplevel "local" optional.  (Strictly speaking,
3511    signature definitions within "local" are not legal SML.)
3512    
3513    Other than that: updates to INSTALL and cm/TODO.
3514    
3515    ----------------------------------------------------------------------
3516    Name: Matthias Blume
3517    Date: 2002/02/08 15:00:00 EST
3518    Tag: blume-20020208-uniquepid
3519    Description:
3520    
3521    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3522    
3523    1. The installer (config/install.sh) has gotten smarter:
3524    
3525         - Configuration options are a bit easier to specify now
3526           (in config/targets).
3527         - Bug in recognizing .tar.bz2 files fixed.
3528         - Installer automatically resolves dependencies between
3529           configuration options (e.g., if you ask for eXene, you will
3530           also get cml -- regardless whether you asked for it or not).
3531         - Installer can run in "quieter mode" by setting the environment
3532           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3533           "completely silent", though.
3534         - Build HashCons library as part of smlnj-lib.
3535    
3536    2. A new scheme for assigning persistent identifiers to compilation
3537       units (and, by extension, to types etc.) has been put into place.
3538       This fixes a long-standing bug where types and even dynamic values
3539       can get internally confused, thereby compromising type safety
3540       (abstraction) and dynamic correctness.  See
3541    
3542         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3543    
3544       for an example of how things could go wrong until now.
3545    
3546       The downside of the new scheme is that pids are not quite as
3547       persistent as they used to be: CM will generate a fresh pid
3548       for every compilation unit that it thinks it sees for the first
3549       time.  That means that if you compile starting from a clean, fresh
3550       source tree at two different times, you end up with different
3551       binaries.
3552    
3553       Cutoff recompilation, however, has not been compromised because
3554       CM keeps pid information in special caches between runs.
3555    
3556    ----------------------------------------------------------------------
3557    Name: Lal George
3558    Date: 2002/02/07 15:34:13 EST 2002
3559    Tag: <none>
3560    Description:
3561    
3562    Compilers that generate assembly code may produce  global labels
3563    whose value is resolved at link time. The various peephole optimization
3564    modules did not take this in account.
3565    
3566    TODO. The Labels.addrOf function should really return an option
3567    type so that clients are forced to deal with this issue, rather
3568    than an exception being raised.
3569    
3570    ----------------------------------------------------------------------
3571    Name: Lal George
3572    Date: 2002/02/06 13:55:02 EST
3573    Tag: george-20020206-ra-breakup
3574    Description:
3575    
3576    1. A bug fix from Allen.
3577    
3578        A typo causes extra fstp %st(0)'s to be generated at compensation
3579        edges, which might cause stack underflow traps at runtime.  This
3580        occurs in fft where there are extraneous fstps right before the 'into'
3581        trap instruction (in this case they are harmless since none of the
3582        integers overflow.)
3583    
3584    2. Pulled out various utility modules that were embedded in the modules
3585       of the register allocator. I need these modules for other purposes, but
3586       they are not complete enough to put into a library (just yet).
3587    ----------------------------------------------------------------------
3588    Name: Matthias Blume
3589    Date: 2002/01/31 16:05:00 EST
3590    Tag: blume-20020131-sparc-ccalls
3591    Description:
3592    
3593    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3594       of extra stack space by mistake.  Fixed.
3595    
3596    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3597    
3598    ----------------------------------------------------------------------
3599    Name: Allen Leung
3600    Date: 2002/01/30
3601    Tag: leunga-20020130-nowhere-bug-fix
3602    Description:
3603    
3604       MLRISC bug fixes:
3605       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3606       2. MachineInt.fromString was negating its value.
3607    
3608    ----------------------------------------------------------------------
3609    Name: Matthias Blume
3610    Date: 2002/01/29
3611    Tag: blume-20020129-INSTALL
3612    Description:
3613    
3614    - Added somewhat detailed installation instructions (file INSTALL).
3615    - Fixed curl-detection bug in config/install.sh.
3616    - It is now possible to select the URL getter using the URLGETTER
3617      environment variable:
3618    
3619          not set / "unknown"      --> automatic detection (script tries wget,
3620                                       curl, and lynx)
3621          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3622                                       how to properly invoke them)
3623          other                    --> use $URLGETTER directly, it must take
3624                                       precisely two command-line arguments
3625                                       (source URL and destination file name)
3626    
3627    ----------------------------------------------------------------------
3628    Name: Matthias Blume
3629    Date: 2002/01/28
3630    Tag: blume-20020128-sparc-ccalls
3631    Description:
3632    
3633    - Fixed problem with calculation of "used" registers in sparc-c-calls.
3634    - Make use of the allocParam argument in sparc-c-calls.
3635    
3636    ----------------------------------------------------------------------
3637    Name: Matthias Blume
3638    Date: 2002/01/28
3639    Tag: blume-20020128-allocParam
3640    Description:
3641    
3642    John Reppy:  Changes c-calls API to accept client-callback for
3643    allocating extra stack space.
3644    me: Corresponding changes to mlriscGen (using a dummy argument that
3645        does not change the current behavior).
3646    
3647    ----------------------------------------------------------------------
3648    Name: Matthias Blume
3649    Date: 2002/01/28 12:00:00
3650    Tag: Release_110_38
3651    Description:
3652    
3653    This time for real!!!
3654    
3655    ----------------------------------------------------------------------
3656    Name: Matthias Blume
3657    Date: 2002/01/28 10:56:00 EST
3658    Tag: blume-20020128-retraction
3659    Description:
3660    
3661    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
3662       with blume-Release_110_38-retracted.)
3663    
3664    1. Fixed a problem with incorrect rounding modes in real64.sml.
3665       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
3666    
3667    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
3668       and enums fixed.  The naming of corresponding ML identifiers should
3669       now be consistent again.
3670    
3671    ----------------------------------------------------------------------
3672    Name: Allen Leung
3673    Date: 2002/01/27
3674    Tag: leunga-20020127-nowhere
3675    Description:
3676    
3677       Added a target called nowhere in the configuration scripts.
3678       Enabling this will build the MLRISC 'nowhere' tool (for translating
3679       programs with where-clauses into legal SML code) during installation.
3680    
3681    ----------------------------------------------------------------------
3682    Name: Matthias Blume
3683    Date: 2002/01/25 21:27:00 EST
3684    Tag: blume-Release_110_38-retracted
3685    Description:
3686    
3687    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
3688    
3689    README will be added later.
3690    
3691    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
3692    (Release_110_38) removed.  Reason: Last-minute bug fixes.
3693    
3694    ----------------------------------------------------------------------
3695    Name: Matthias Blume
3696    Date: 2002/01/25
3697    Tag: blume-20020125-ffi
3698    Description:
3699    
3700    A large number of tweaks and improvements to ml-nlffi-lib and
3701    ml-nlffigen:
3702    
3703       - ML represenation types have been streamlined
3704       - getter and setter functions work with concrete values, not abstract
3705         ones where possible
3706       - ml-nlffigen command line more flexible (see README file there)
3707       - some bugs have been fixed (hopefully)
3708    
3709    ----------------------------------------------------------------------
3710    Name: Lal George
3711    Date: 2002/01/24
3712    Tag: george-20020124-risc-ra-interface
3713    Description:
3714    
3715       There is a dramatic simplification in the interface to the
3716       register allocator for RISC architectures as a result of making
3717       parallel copy instructions explicit.
3718    
3719    ----------------------------------------------------------------------
3720    Name: Matthias Blume
3721    Date: 2002/01/22
3722    Tag: blume-20020122-x86-ccalls
3723    Description:
3724    
3725    Bug fix for c-calls on x86 (having to do with how char- and
3726    short-arguments are being handled).
3727    
3728    ----------------------------------------------------------------------
3729    Name: Matthias Blume
3730    Date: 2002/01/21
3731    Tag: blume-20020121-ff
3732    Description:
3733    
3734    Another day of fiddling with the FFI...
3735    
3736    1. Bug fix/workaround:  CKIT does not complain about negative array
3737       dimensions, so ml-nlffigen has to guard itself against this possibility.
3738       (Otherwise a negative dimension would send it into an infinite loop.)
3739    
3740    2. Some of the abstract types (light objects, light pointers, most "base"
3741       types) in structure C are now eqtypes.
3742    
3743    3. Added constructors and test functions for NULL function pointers.
3744    
3745    ----------------------------------------------------------------------
3746    Name: Matthias Blume
3747    Date: 2002/01/18
3748    Tag: blume-20020118-ready-for-new-release
3749    Description:
3750    
3751    Made config/srcarchiveurl point to a new place.  (Will provide boot
3752    files shortly.)
3753    
3754    Maybe we christen this to be 110.38?
3755    
3756    ----------------------------------------------------------------------
3757    Name: Matthias Blume
3758    Date: 2002/01/18
3759    Tag: blume-20020118-more-ffifiddle
3760    Description:
3761    
3762    Today's FFI fiddling:
3763    
3764      - Provided a structure CGetSet with "convenient" versions of C.Get.* and
3765        C.Set.* that use concrete (MLRep.*) arguments and results instead
3766        of abstract ones.
3767    
3768      - Provided word-style bit operations etc. for "int" representation
3769        types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
3770        and Long.
3771    
3772    ----------------------------------------------------------------------
3773    Name: Matthias Blume
3774    Date: 2002/01/18
3775    Tag: blume-20020118-use-x86-fp
3776    Description:
3777    
3778    Now that x86-fast-fp seems to be working, I turned it back on again
3779    by default.  (Seems to work fine now, even with the FFI.)
3780    
3781    Other than that, I added some documentation about the FFI to
3782    src/ml-nlffigen/README and updated the FFI test examples in
3783    src/ml-nlffi-lib/Tests/*.
3784    
3785    ----------------------------------------------------------------------
3786    Name: Allen Leung
3787    Date: 2002/01/17
3788    Tag: leunga-20020117-x86-fast-fp-call
3789    Description:
3790    
3791       1. Fixed a problem with handling return fp values when x86's fast fp
3792          mode is turned on.
3793    
3794       2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
3795          of %f32.
3796    
3797       3. Added a constructor INT32lit to the ast of MLRISC tools.
3798    
3799    ----------------------------------------------------------------------
3800    Name: Matthias Blume
3801    Date: 2002/01/16
3802    Tag: blume-20020116-ffifiddle
3803    Description:
3804    
3805    More fiddling with the FFI interface:
3806    
3807     - Make constness 'c instead of rw wherever possible.  This eliminates
3808       the need for certain explicit coercions.  (However, due to ML's
3809       value polymorphism, there will still be many cases where explicit
3810       coercions are necessary.  Phantom types are not the whole answer
3811       to modeling a subtyping relationship in ML.)
3812    
3813     - ro/rw coersions for pointers added.  (Avoids the detour through */&.)
3814    
3815     - "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
3816       clumsy workaround for varargs problem.)
3817    
3818    ----------------------------------------------------------------------
3819    Name: Lal George
3820    Date: 2002/01/15
3821    Tag: <none>
3822    Description:
3823    
3824    1. Since COPY instructions are no longer native to the architecture,
3825       a generic functor can be used to implement the expandCopies function.
3826    
3827    2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
3828       TEXT segment.
3829    
3830    ----------------------------------------------------------------------
3831    Name: Matthias Blume
3832    Date: 2002/01/15
3833    Tag: blume-20020115-ffiupdates
3834    Description:
3835    
3836    1. Fix for bug resulting in single-precision float values being returned
3837       incorrectly from FFI calls.
3838    
3839    2. Small modifications to C FFI API:
3840    
3841        - memory-allocation routines return straight objects (no options)
3842          and raise an exception in out-of-memory situations
3843        - unsafe extensions to cast between function pointers and pointers
3844          from/to ints
3845        - added structure C_Debug as an alternative to structure C where
3846          pointer-dereferencing (|*| and |*!) always check for null-pointers
3847        - added open_lib' to DynLinkage;  open_lib' works like open_lib
3848          but also takes a (possibly empty) list of existing library handles
3849          that the current library depends on
3850    
3851    ----------------------------------------------------------------------
3852    Name: Matthias Blume
3853    Date: 2002/01/10
3854