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 976, Wed Nov 14 14:54:03 2001 UTC revision 1791, Tue May 10 02:56:53 2005 UTC
# Line 8  Line 8 
8  The form of an entry should be:  The form of an entry should be:
9    
10  Name:  Name:
11  Date:  Date: yyyy/mm/dd
12  Tag: <post-commit CVS tag>  Tag: <post-commit CVS tag>
13  Description:  Description:
14    
15  ----------------------------------------------------------------------  ----------------------------------------------------------------------
16    Name: Matthias Blume (blume (at) tti - c (dot) org)
17    Date: 2005/05/09 21:55:00 CDT
18    Tag: blume-20050509-word64
19    Description:
20    
21    Added a hack to the existing hack known as Word64 to make fromString
22    behave correctly.  I am still not sure whether Word64.scan will work
23    as specified with respect to the interaction of radix and prefix.
24    
25    ----------------------------------------------------------------------
26    Name: Allen Leung (leunga (at) reservoir (dot) com)
27    Date: 2005/05/04 11:50:00 EST
28    Tag: leunga-20050504-checkgc
29    Description:
30    
31    Added a gc protocol checking phase.  This phase is enabled with
32    the flag "check-gc".  "debug-check-gc" turns on the verbose mode.
33    
34    ----------------------------------------------------------------------
35    Name: Matthias Blume (blume (at) tti - c (dot) org)
36    Date: 2005/05/04 10:45:00 CDT
37    Tag: blume-20050504-intinf
38    Description:
39    
40    Fixed a bug in the implementation of div and mod for IntInf.
41    Thanks to Neophytos Michael for reporting the problem.
42    
43    ----------------------------------------------------------------------
44    Name: Matthias Blume (blume (at) tti - c (dot) org)
45    Date: 2005/05/04 10:35:00 CDT
46    Tag: blume-20050504-join
47    Description:
48    
49    Added a "join" combinator to the ParserComb module in smlnj-lib.cm.
50    
51    ----------------------------------------------------------------------
52    Name: Matthias Blume (blume (at) tti - c (dot) org)
53    Date: 2005/02/28 23:40:00 CST
54    Tag: blume-20050228-mVar
55    Description:
56    
57    Fixed serious bug (brown paper bag variety) in new implementation of
58    structure Atom in CML.  (I had accidentally used a mailbox instead of
59    an mvar, leaving the door open for races.)
60    
61    ----------------------------------------------------------------------
62    Name: Matthias Blume (blume (at) tti - c (dot) org)
63    Date: 2005/02/25 15:00:00 CST
64    Tag: Release_110_53
65    Description:
66    
67    New working version (110.53).  NEW BOOTFILES!
68    
69    ----------------------------------------------------------------------
70    Name: Matthias Blume (blume (at) tti - c (dot) org)
71    Date: 2005/02/25 14:50:00 CST
72    Tag: blume-20050225-susp
73    Description:
74    
75    Brought back SMLofNJ.Susp.  The underlying suspension type is the one
76    implemented in Core -- which means that it is the same as the one used
77    by the lazy extension.
78    
79    ----------------------------------------------------------------------
80    Name: Matthias Blume (blume (at) tti - c (dot) org)
81    Date: 2005/02/24 16:50:00 CST
82    Tag: blume-20050224-cml-atom
83    Description:
84    
85    Simpler and at the same time more general implementation of structure
86    Atom in CML.
87    
88    ----------------------------------------------------------------------
89    Name: Matthias Blume (blume (at) tti - c (dot) org)
90    Date: 2005/02/15 17:35:00 CST
91    Tag: blume-20050215-tools
92    Description:
93    
94    Created new "tools" directory under "src" and moved "TraceDebugProf"
95    there.
96    
97    ----------------------------------------------------------------------
98    Name: Matthias Blume (blume (at) tti - c (dot) org)
99    Date: 2005/02/10 17:55:00 CST
100    Tag: blume-20050210-longlong
101    Description:
102    
103    Implemented "long long" arguments and results for NLFFI.  (Only the
104    PPC/MacOS implementation is complete, the other backends still need to
105    be updated.)
106    
107    ----------------------------------------------------------------------
108    Name: Matthias Blume (blume (at) tti - c (dot) org)
109    Date: 2005/01/24 17:40:00 CST
110    Tag: blume-20050124-mlyacc
111    Description:
112    
113    Minor cleanup in ML-Yacc rule printing mechanism.  This should fix a
114    problem with certain "as" patterns which previously got rendered
115    using incorrect syntax.
116    
117    ----------------------------------------------------------------------
118    Name: Matthias Blume (blume (at) tti - c (dot) org)
119    Date: 2005/01/18 12:00:00 CST
120    Tag: blume-20050118-profile
121    Description:
122    
123    Made time profiling code (interrupt handler) in runtime system aware
124    of new array representation.
125    
126    ----------------------------------------------------------------------
127    Name: Matthias Blume (blume (at) tti - c (dot) org)
128    Date: 2005/01/14 18:00:00 CST
129    Tag: blume-20050114-heap2exec
130    Description:
131    
132    Implemented new (but still experimental) heap2exec facility.  This is
133    tested under Mac OS X and should work under Linux (will test shortly).
134    It will probably also work on the Sparc (will test some time later).
135    
136      - removed old "HACKED_STANDALONE" hack from runtime
137    
138    To be able to test this, uncomment the request for "heap2asm" in
139    config/targets prior to installation.  (Notice that this is different
140    from "heap2exec" mentioned below.  Not a typo.)
141    
142    To perform an actual test, run the command
143    
144       $ bin/heap2exec heapfile execfile
145    
146    (You can put heap2exec on your shell's path.)
147    
148    For example, run
149    
150      $ bin/heap2exec bin/.heap/ml-yacc.ppc-darwin mly
151    
152    This will create a standalone executable called "mly" which you can
153    then invoke directly as a command.
154    
155    ----------------------------------------------------------------------
156    Name: Matthias Blume (blume (at) tti - c (dot) org)
157    Date: 2005/01/07 11:44:00 CST
158    Tag: blume-20050107-mlstring
159    Description:
160    
161    fixed off-by-one error in ML_STRING macro (globals.c)
162    
163    ----------------------------------------------------------------------
164    Name: Matthias Blume (blume (at) tti - c (dot) org)
165    Date: 2004/12/23 18:00:00 CST
166    Tag: blume-20041223-santa
167    Description:
168    
169    Made ml-build script "smarter" (but only very little).
170    
171    ----------------------------------------------------------------------
172    Name: Matthias Blume (blume (at) tti - c (dot) org)
173    Date: 2004/12/21 15:05:00 CST
174    Tag: blume-20041221-longlong
175    Description:
176    
177    * Implemented access to signed and unsigned long long data in NLFFI.
178    (The parameter-passing part of the picture has not complete. But
179    data structure access seems to work.)
180    
181    * Fixed CM's incorrect assumption that the PPC is little-endian.
182      (On the Mac, it is big-endian.  And that's currently our only
183      PPC platform.)
184    
185    ----------------------------------------------------------------------
186    Name: Matthias Blume (blume (at) tti - c (dot) org)
187    Date: 2004/12/21 12:50:00 CST
188    Tag: blume-20041221-memory
189    Description:
190    
191    Some cleanup in the $c/memory.cm library: separated some concerns by
192    moving allocation code and memory access code each into their own
193    files.
194    
195    ----------------------------------------------------------------------
196    Name: Allen Leung (leunga (at) reservoir (dot) com)
197    Date: 2004/12/17 16:12:00 EST
198    Tag: leunga-20041217-cygwin-smlnj-home
199    Description:
200    
201        The Unix I/O library of SML/NJ on cygwin does not understand
202    Windows style pathname, so problems arise when SMLNJ_HOME is set
203    to a Windows style pathname.  _run-sml now converts SMLNJ_HOME
204    to a POSIX pathname on cygwin.
205    
206    ----------------------------------------------------------------------
207    Name: Matthias Blume (blume (at) tti - c (dot) org)
208    Date: 2004/12/16 13:35:00 CST
209    Tag: Release_110_52
210    Description:
211    
212    Last-minute changes incorporated into 110.52.  Release tag moved.
213    
214    The changes:
215       - HashString.hashString' -> HashString.hashSubstring
216       - bug fix in UnivariateStats
217    
218    ----------------------------------------------------------------------
219    Name: Matthias Blume (blume (at) tti - c (dot) org)
220    Date: 2004/12/15 23:40:00 CST
221    Tag: blume-20041215-hashSubstring
222    Description:
223    
224    - HashString.hashString' -> HashString.hashSubstring
225    - corresponding changes in atom.sml
226    - "de-compressed" (aka. un-obfuscated) code for UnivariateStats and
227      added some comments
228    
229    ----------------------------------------------------------------------
230    Name: Matthias Blume (blume (at) tti - c (dot) org)
231    Date: 2004/12/15 15:30:00 CST
232    Tag: (Release_110_52)
233    Description:
234    
235    New working version (110.52).  NEW BOOTFILES!
236    
237    ----------------------------------------------------------------------
238    Name: Matthias Blume (blume (at) tti - c (dot) org)
239    Date: 2004/12/15 12:45:00 CST
240    Tag: blume-20041215-spaces
241    Description:
242    
243    More on the space problem (this time for Win32).
244    
245    ----------------------------------------------------------------------
246    Name: Matthias Blume (blume (at) tti - c (dot) org)
247    Date: 2004/12/14 17:30:00 CST
248    Tag: blume-20041214-spaces
249    Description:
250    
251    Hacked some of the scripts (in particular: the installer) to cope with
252    spaces in filenames a bit better.  But beware: the current "solution"
253    is likely still full of bugs and inherently incomplete.  (We need to
254    do away with those shell scripts for a comprehensive solution.)
255    
256    ----------------------------------------------------------------------
257    Name: Matthias Blume (blume (at) tti - c (dot) org)
258    Date: 2004/12/13 14:45:00 CST
259    Tag: blume-20041213-ml-makedepend
260    Description:
261    
262    Fixed bug in code for ml-makedepend.
263    
264    ----------------------------------------------------------------------
265    Name: Matthias Blume (blume (at) tti - c (dot) org)
266    Date: 2004/12/09 16:30:00 CST
267    Tag: blume-20041209-statistics
268    Description:
269    
270    Added two simple but potentially useful statistics modules to SML/NJ Library.
271    (See CHANGES file there.)
272    
273    ----------------------------------------------------------------------
274    Name: Matthias Blume (blume (at) tti - c (dot) org)
275    Date: 2004/12/01 16:50:00 CST
276    Tag: blume-20041201-atom
277    Description:
278    
279    smlnj-lib:
280    
281      Added function HashString.hashString' for substrings.
282      Hand-inlined CharVector.foldl into HashString (for speed).
283      Modified implementation of structure Atom to avoid extracting
284      strings from substrings unless necessary.
285      (Also see CHANGES file for smlnj-lib.)
286    
287    ----------------------------------------------------------------------
288    Name: Matthias Blume (blume (at) tti - c (dot) org)
289    Date: 2004/11/24 22:15:00 CST
290    Tag: blume-20041124-cml
291    Description:
292    
293    Made sure CML compiles when Position = Int64.
294    
295    ----------------------------------------------------------------------
296    Name: Matthias Blume (blume (at) tti - c (dot) org)
297    Date: 2004/11/24 14:30:00 CST
298    Tag: blume-20041124-position
299    Description:
300    
301    The compiler can now be compiled in a mode that makes structure Position
302    equal to Int64.  The default, however, is unchanged (Position = Int31)
303    for the time being.
304    
305    To enable 64-bit positions, use the following procedure:
306    
307    1. Start sml
308    2. Autoload $smlnj/cmb.cm (if not already autoloaded)
309    3. Type
310          #set (CMB.symval "USE_64_BIT_POSITIONS") (SOME 1);
311    4. Run CMB.make() as usual.
312    
313    This is barely tested.  The only test so far was a little SML program
314    counting the number of characters in an 8-gigabyte file by
315    reading it character-by-character.  That test was successful.
316    
317    In support of 64-bit positions, a number of new functions have been
318    added to the runtime system.
319    
320    ----------------------------------------------------------------------
321    Name: Matthias Blume (blume (at) tti - c (dot) org)
322    Date: 2004/11/23 14:45:00 CST
323    Tag: blume-20041123-useFile
324    Description:
325    
326    Fixed a problem with unhelpful error messages related to problems with
327    .cm- or .sml files that appear as part of the sml command line.
328    
329    ----------------------------------------------------------------------
330    Name: Matthias Blume (blume (at) tti - c (dot) org)
331    Date: 2004/11/18 15:40:00 CST
332    Tag: Release_110_51
333    Description:
334    
335    New working version (110.51).  NEW BOOTFILES!
336    
337    ----------------------------------------------------------------------
338    Name: Matthias Blume (blume (at) tti - c (dot) org)
339    Date: 2004/11/18 15:35:00 CST
340    Tag: <none>
341    Description:
342    
343    Enabled dlopen and friends for FreeBSD (as recommended by Johannes 5
344    Joemann).
345    
346    ----------------------------------------------------------------------
347    Name: Allen Leung (leunga (at) reservoir (dot) com)
348    Date: 2004/11/17 16:05:21 EST 2004
349    Tag: leunga-20041117-mlrisc-live-kill
350    Description:
351    
352       Added support for MLTree constructs LIVE and KILL
353    to all the architectures.
354    
355    ----------------------------------------------------------------------
356    Name: Matthias Blume (blume (at) tti - c (dot) org)
357    Date: 2004/11/13 00:20:00 CST
358    Tag: blume-20041113-versiontool
359    Description:
360    
361    - Stripped down the versiontool: It now only handles the version number.
362      The date string is generated at bootstrap time (during makeml).
363    
364    - In a previous commit, fixed a minor issue with how polyequal is being
365      translated.  In particular, the code now "looks through" abstractions.
366      This results in slightly fewer polyEqual warnings and hopefully slightly
367      more efficient code.  Important examples for where this matters are
368      the new int64 and word64 types.
369    
370    ----------------------------------------------------------------------
371    Name: Matthias Blume (blume (at) tti - c (dot) org)
372    Date: 2004/11/12 00:30:00 CST
373    Tag: blume-20041112-int64
374    Description:
375    
376    Structure Int64 fully hooked in.  (The implementation is not very
377    efficient, though.)
378    
379    ----------------------------------------------------------------------
380    Name: Matthias Blume (blume (at) tti - c (dot) org)
381    Date: 2004/11/11 17:30:00 CST
382    Tag: blume-20041111-more64
383    Description:
384    
385    All the pieces of Word64 are now there, with the exception of the
386    conversions from and to LargeWord. (Eventually these need to be identities,
387    but for the time being they don't even make sense because LargeWord is
388    32-bit wide.)
389    
390    Also started to add similar support for Int64, but major pieces of that
391    are still missing.
392    
393    ----------------------------------------------------------------------
394    Name: Matthias Blume (blume (at) tti - c (dot) org)
395    Date: 2004/11/11 00:15:00 CST
396    Tag: blume-20041111-word64
397    Description:
398    
399    Structure Word64 is now (almost) complete, word literals and patterns
400    seem to work.  There are a few odd pieces missing.  In particular,
401    I didn't do the {from,to}LargeWord parts because LargeWord is still Word32
402    at the moment.
403    
404    Making Word64 official would mean that LargeWord becomes Word64.  But
405    this requires extreme care because most word-word conversions have to
406    go through LargeWord, so making a mistake means loss of efficiency or
407    worse.  Eventually there will be a solution similar to (but actually
408    simpler than) what I did with IntInf.
409    
410    ----------------------------------------------------------------------
411    Name: Matthias Blume (blume (at) tti - c (dot) org)
412    Date: 2004/11/10 18:12:00 CST
413    Tag: blume-20041110-64bit
414    Description:
415    
416    More 64-bit hacking (but still not even half-way there yet).
417    Also, some assorted improvements to the handling of 8-bit words.
418    
419    ----------------------------------------------------------------------
420    Name: Matthias Blume (blume (at) tti - c (dot) org)
421    Date: 2004/11/09 17:50:00 CST
422    Tag: <none>
423    Description:
424    
425    Started adding some infrastructure for supporting 64-bit int- and
426    word-types.  (Still in its very early stages.)
427    
428    ----------------------------------------------------------------------
429    Name: Matthias Blume (blume (at) tti - c (dot) org)
430    Date: 2004/10/28 10:45:00 CDT
431    Tag: Release_110_50
432    Description:
433    
434    New working version (110.50).  NEW BOOTFILES!
435    
436                 =====================
437    
438    Also:
439    
440      - Changend config/srcarchiveurl from a file just
441        containing the URL string into a file containing
442        shell script code.  The code has access to the $VERSION variable.
443      - Made corresponding changes to config/install.sh and config/unpack.
444      - Default contents of config/srcarchiveurl uses $VERSION and
445        normally does not have to be edited to reflect a version change.
446    
447      (As a result, a version change can be done by just editing
448       config/version, the rest is now automatic.)
449    
450    ----------------------------------------------------------------------
451    Name: Matthias Blume (blume (at) tti - c (dot) org)
452    Date: 2004/10/27 17:50:00 CDT
453    Tag: blume-20041027-btrace-msg
454    Description:
455    
456    BackTrace.monitor now also reports the source of the exception that
457    triggered the trace.
458    
459    ----------------------------------------------------------------------
460    Name: Matthias Blume (blume (at) tti - c (dot) org)
461    Date: 2004/10/27 17:20:00 CDT
462    Tag: blume-20041027-x86-c-calls
463    Description:
464    
465    This is the HISTORY entry for two earlier commits, both concerning
466    the x86 c-calls code in MLRISC:
467    
468      - added a missing LOAD in the code that deals with struct arguments
469      - made sure the caller does not add the wrong number of bytes to the
470        stack pointer after a call of a function returning a struct
471        (the callee already pops the implicit argument which points to
472         the space reserved for the result)
473    
474    ----------------------------------------------------------------------
475    Name: Allen Leung (leunga (at) reservoir (dot) com)
476    Date: 2004/10/24 14:00:00 EST
477    Tag: leunga-20041024-x86-gas-fucomip
478    Description:
479    
480       John discovered a bug in the syntax of fucomip.
481       The opcodes FU?COMIP? have been changed to
482    
483           fu?comip? %st(i), %st
484    
485    ----------------------------------------------------------------------
486    Name: Matthias Blume (blume (at) tti - c (dot) org)
487    Date: 2004/10/20 15:06:00 CDT
488    Tag: blume-20041020-standalone-backtrace
489    Description:
490    
491    Added a mechanism for getting back-trace information from standalone
492    programs.  Here is how it works:
493    
494       1. The part of the program from which you want to get backtrace
495          information (usually the whole program) should be wrapped with
496          BackTrace.monitor.  This is a (unit->'a)->'a function, and your
497          main program could be modified from something like
498    
499             fun main (pgm, args) = ...
500    
501         to
502    
503            fun main (pgm, args) = BackTrace.monitor (fn () => ...)
504    
505       2. To be able to access BackTrace.monitor, you have to add
506          library $smlnj-tdp/plugins.cm to the .cm file that contains your
507          main function.
508    
509       3. Remove all compiled code (i.e., all the .cm/ subdirectories that
510          CM might have created in the past for your project).
511    
512       4. Build the system using this command line:
513    
514             ml-build -Ctdp.instrument=true \$smlnj-tdp/back-trace.cm \
515                      myprog.cm MyProg.main myprog
516    
517          instead of the usual
518    
519             ml-build myprog.cm MyProg.main myprog
520    
521    I changed a library name:
522    
523       $/trace-debug-profile.cm --> $smlnj-tdp/plugins.cm
524    
525    New libraries:
526    
527       $smlnj-tdp/back-trace.cm
528          -- when loaded causes the back-trace plugin to be installed
529       $smlnj-tdp/coverage.cm
530          -- when loaded causes the coverage plugin to be installed
531    
532    ----------------------------------------------------------------------
533    Name: Matthias Blume (blume (at) tti - c (dot) org)
534    Date: 2004/10/18 16:45:00 CDT
535    Tag: blume-20041018-groupowner
536    Description:
537    
538    Added an "obsolete" warning for the "group owner" syntax to CM's parser.
539    
540    Eliminated group owner specs from .cm files throughout the source tree.
541    
542    ----------------------------------------------------------------------
543    Name: Matthias Blume (blume (at) tti - c (dot) org)
544    Date: 2004/10/15 15:45:00 CDT
545    Tag: blume-20041015-coverage
546    Description:
547    
548    * Test coverage tool added!
549    * Further reorganization of tracing-, debugging-, and profiling support:
550    
551        - moved original BTImp -- now called BackTrace -- into a separate
552          library called $/trace-debug-profile.cm
553        - eliminated all mentions of BTrace from SMLofNJ.Internals
554        - only the instrumentation mechanism is now left in the compiler proper
555        - BackTrace module is a plugin which is NOT plugged in by default
556        - Coverage module is another such plugin
557    
558    To get the benefits of any of these plugin modules, the code in
559    question must be compiled with tdp instrumentation turned on.  This
560    can be done by setting SMLofNJ.Internals.TDP.mode to true.
561    (The ref cell is also controlled via the -Ctdp.instrument=... switch.)
562    
563    Plugins are selected at link time.  (Pre-compiled instrumented code
564    can be re-loaded with different plugins in effect.)  When an
565    instrumented module is linked, whatever plugins are at that time
566    enabled will come into effect for that module.
567    
568    To enable the back-trace plugin, load library $/trace-debug-profile.cm
569    and invoke BackTrace.install() (e.g., from the interactive prompt).
570    To enable the coverage plugin, load the same library and invoke
571    Coverage.install().
572    
573    Back-traces are generated automatically on uncaught exceptions and
574    when the code in question explicitly invokes BackTrace.trigger().
575    
576    Coverage (and execution frequency-) information must be queried
577    explicitly by calling Coverage.not_covered and Coverage.hot_spots.
578    
579    ----------------------------------------------------------------------
580    Name: Matthias Blume (blume (at) tti - c (dot) org)
581    Date: 2004/10/14 17:40:00 CDT
582    Tag: blume-20041014-tdp-core
583    Description:
584    
585    Snapshot of a significant overhaul of how the trace/debug/profile support
586    is hooked into the system (specifically: Core and SMLofNJ.Internals).
587    
588    ----------------------------------------------------------------------
589    Name: Matthias Blume (blume (at) tti - c (dot) org)
590    Date: 2004/10/13 16:34:00 CDT
591    Tag: blume-20041013-tdp
592    Description:
593    
594    Some rationalization of names:
595    
596       structure BTrace -> structure TDPInstrument
597       etc.
598    
599    This is is preparation of using the original back-trace
600    instrumentation for other purposes.  "TDP" stands for
601    Trace/Debug/Profile.
602    
603    The control flag controlling whether instrumentation is on or off is now
604    registered under a different name, so instead of running sml as
605    
606     sml -Cinstrument.btrace-mode=true
607    
608    one has to say
609    
610     sml -Ctdp.instrument=true
611    
612    ----------------------------------------------------------------------
613    Name: Matthias Blume (blume (at) tti - c (dot) org)
614    Date: 2004/10/11 16:37:00 CDT
615    Tag: blume-20041011-regions
616    Description:
617    
618    Made some minor modifications to elabcore.sml to have source regions
619    be propagated more tightly -- resulting in better (i.e., smaller)
620    regions being reported in error- and debug messages.
621    
622    ----------------------------------------------------------------------
623    Name: Matthias Blume (blume (at) tti - c (dot) org)
624    Date: 2004/10/08 22:50:00 CDT
625    Tag: blume-20041008-cmkw
626    Description:
627    
628    Fixed handling of keywords in .cm files: After seeing "is" the lexer
629    treats subsequent occurrences of "group", "library", "source", "is",
630    "*", and "-" as ordinary identifiers rather than keywords.
631    
632    Most seriously, this fixes a problem with CM's "shell" tool.  The tool
633    is supposed to accept a tool argument called "source", but this did
634    not work because of the clash with the keyword.
635    
636    ----------------------------------------------------------------------
637    Name: Matthias Blume (blume (at) tti - c (dot) org)
638    Date: 2004/10/07 16:00:00 CDT
639    Tag: blume-20041007-cleanup
640    Description:
641    
642    Assorted cleanup work:
643    
644      - got rid of intstrmap in favor of using the library's
645        hash table implementation
646      - threw out most of the pathnames stuff, as it was not used anyway
647      - simplified tokentable implementation
648      - fixed some minor spelling errors
649    
650    ----------------------------------------------------------------------
651    Name: Matthias Blume (blume (at) tti - c (dot) org)
652    Date: 2004/10/06 15:15:15 CDT
653    Tag: blume-20041006-handler
654    Description:
655    
656    Cleaned up the absyn to reflect the invariant that HANDLE always
657    carries a FNexp as part of the type definition.  This eliminates some
658    superfluous sanity checks at runtime down the road.
659    
660    Some minor cleanup of the btrace code.
661    
662    ----------------------------------------------------------------------
663    Name: Matthias Blume (blume (at) tti - c (dot) org)
664    Date: 2004/10/01 10:20:30 CDT
665    Tag: blume-20041001-slave
666    Description:
667    
668    Added hack to make slave mode work in the presence of the version
669    tool.  (Still, since the master does two passes over the code for
670    CMB.make, the release number gets bumped twice when slaves are
671    attached.  I don't know if this is worth fixing...)
672    
673    ----------------------------------------------------------------------
674    Name: Matthias Blume (blume (at) tti - c (dot) org)
675    Date: 2004/09/30 10:55:00 CDT
676    Tag: blume-20040930-version
677    Description:
678    
679    * Moved the "version" magic into its own little library under
680      src/system/smlnj/internal.  This avoids expensive reconstruction of
681      a stable src/compiler/core.cm.
682    
683    * At the same time, structure CompilerVersion is now known as
684      structure SMLNJVersion.
685    
686    * Arranged for the version tool to NOT kick in when rebuilding the system
687      (makeml -rebuild, fixpt).  Otherwise one would never reach a fixpoint.
688      Also, loading the versiontool does not work when rebuilding the system
689      because CM is not properly initialized at that time.
690    
691    ----------------------------------------------------------------------
692    Name: Matthias Blume (blume (at) tti - c (dot) org)
693    Date: 2004/09/29 14:00:00 CDT
694    Tag: blume-20040929-autoversion
695    Description:
696    
697    Implemented some CM magic to have
698        file src/compiler/TopLevel/main/version.sml
699    generated automagically.
700    The version is taken from two files: config/version and config/release.
701    The first is expected to contain a two-part version number such as 110.49.
702    The second should contain a single number, but it may be missing.
703    
704    If the environment variable VERSIONTOOL_BUMP_RELEASE is defined at the
705    time the version tool is loaded (which is the first time you say CMB.make),
706    then the tool will increment the value stored in config/release every
707    time CMB.make is invoked.
708    
709    The binfile format is now insensitive to anything beyond the first
710    two components of a version number, so bumping the release does not render
711    binfiles incompatible.  Auto-bumping can be used to keep track of versions
712    during development without invalidating existing binfiles.
713    
714    In any case, every CMB.make updates the date information in version.sml.
715    (This is the date that is printed in the banner.)
716    
717    ----------------------------------------------------------------------
718    Name: Matthias Blume (blume (at) tti - c (dot) org)
719    Date: 2004/09/28 10:53:00 CDT
720    Tag: blume-20040928-controls
721    Description:
722    
723    Some cleanup of the controls code.
724    
725    ----------------------------------------------------------------------
726    Name: Matthias Blume (blume (at) tti - c (dot) org)
727    Date: 2004/09/27 22:08:00 CDT
728    Tag: blume-20040927-controls
729    Description:
730    
731    Added two pieces of functionality to the Controls interface:
732    
733    1.   val save'restore: 'a control -> unit -> unit
734    
735       grabs the current value of the control in stage 1 and restores it
736       in stage 2.
737    
738    2.   val set' : 'a control * 'a -> unit -> unit
739    
740       stores the given value into the control in stage 2 (i.e., delayed)
741       but does all error checking in stage 1.
742       (This is for string controls that need to do parse their argument --
743        something that might fail.  In some cases, notably in CM, one
744        already knows the intended argument but wants to delay the actual
745        assignment until a time when error recovery would be more difficult.)
746    
747    Changed the handling of controls in tool arguments to classes "sml" and
748    "lazysml":
749       - use Controls.save'restore as a more robust way of restoring the
750         old value (in particular: without having to re-parse the string)
751       - use controls to handle the "overload" keyword in the init group
752         (I believe this change actually fixes a long-standing obscure bug.)
753    
754    ----------------------------------------------------------------------
755    Name: Matthias Blume (blume (at) tti - c (dot) org)
756    Date: 2004/09/27 17:00:00 CDT
757    Tag: blume-20040927-lazysml
758    Description:
759    
760    Added a new tool class called "lazysml" to CM's tool chest.  The only
761    difference to "sml" is that compilation is done with Control.lazysml
762    set to true.  A source of class "lazysml" is automatically recognized
763    by a file name suffix of ".lml".
764    
765    In addition to the above feature, the original class "sml" now also
766    supports a tool argument "lazy" which has the same effect.  As a
767    result, the following three lines are equivalent:
768    
769        foo.sml : lazysml
770        foo.sml : sml (lazy)
771        foo.sml (lazy)
772    
773    The setting goes into effect both during parsing and during
774    compilation.  The original setting is restored right after parsing and
775    after compilation, respectively.
776    
777    In addition to all the above, there is also a general mechanism to set
778    ANY of the "controls" that are available at the command line via
779    "-C..." on a per-sml-file basis.  The same rules that apply for "lazy"
780    apply as well. (In fact, "lazy" is implemented as a special case of
781    the general mechanism.)
782    
783    The .cm file syntax uses a new keyword tool argument called "with".
784    There are several ways of indicating the desired settings:
785    
786        foo.sml (with:parser.quotations=true)
787        foo.sml (with:(name:parser.quotations value:true))
788        foo.sml (with:(name:name1 value:value1 name:name2 value:value2 ...))
789        foo.sml (with:(name1=value1 name2=value2 ...))
790        foo.sml (with:(name1=value1 name:name2 value:value2 name3=value3 ...))
791    
792        etc.
793    
794    Another possible abbreviation is to leave out the =v or value:v part
795    if the name refers to a boolean control (in which case the value is
796    taken to be true).  Thus, one could get lazy sml also by saying:
797    
798        foo.sml (with:parser.lazy-keyword=true)
799        foo.sml (with:parser.lazy-keyword)
800        foo.sml (with:(name:parser.lazy-keyword value:true))
801        foo.sml (with:(name:parser.lazy-keyword))
802    
803    ----------------------------------------------------------------------
804    Name: Matthias Blume (blume (at) tti - c (dot) org)
805    Date: 2004/09/24 16:22:00 CDT
806    Tag: blume-20040924-ppc-long-branch
807    Description:
808    
809    Turned message about "emiting long form of branch" off by default.
810    Added a control flag to turn it back on when desired.
811    
812    ----------------------------------------------------------------------
813    Name: Matthias Blume (blume (at) tti - c (dot) org)
814    Date: 2004/09/24 16:05:00 CDT
815    Tag: blume-20040924-rounding
816    Description:
817    
818    Applied patch for setting rounding modes under Mac OS X.  Thanks to
819    Melissa O'Neill for providing the code!
820    
821    ----------------------------------------------------------------------
822    Name: Matthias Blume (blume (at) tti - c (dot) org)
823    Date: 2004/09/23 17:30:00 CDT
824    Tag: blume-20040923-envvars
825    Description:
826    
827    1. Changed definition of type ControlRegistry.registry_tree to
828       include control_info (i.e., the name of the controlling environment
829       variable).
830    
831    2. Added command-line flags -e and -E to print the names of environment
832       variables that can be used to control internal settings.
833       (This uses the new API mentioned in 1.)
834    
835    ----------------------------------------------------------------------
836    Name: Matthias Blume (blume (at) tti - c (dot) org)
837    Date: 2004/09/13 16:50:00 CDT
838    Tag: Release_110_49
839    Description:
840    
841    New working version (110.49).  NEW BOOTFILES!
842    
843    ----------------------------------------------------------------------
844    Name: Matthias Blume
845    Date: 2004/09/13 16:20:00 CDT
846    Tag: blume-20040913-config-mlrisc
847    Description:
848    
849    Put target "mlrisc" back into the default list.
850    (There is no harm in having it, and some users have expressed their
851    wish to have "mlrisc" included by default.)
852    
853    ----------------------------------------------------------------------
854    Name: John Reppy
855    Date: 2004/09/13
856    Tag: jhr-20040913-signals
857    Description:
858    
859    Fixed the signal masking code to properly nest mask/unmask operations
860    on a per-signal basis.
861    
862    ----------------------------------------------------------------------
863    Name: Matthias Blume (blume (at) tti - c (dot) org)
864    Date: 2004/09/08 13:20:00 CDT
865    Tag: blume-20040908-heap-magic
866    Description:
867    
868    Bumped the heap macig number to 0x09082004 to account for the changed
869    layout of the ML frame under MacOS X.
870    
871    ----------------------------------------------------------------------
872    Name: Allen Leung (leunga (at) reservoir (dot) com)
873    Date: 2004/09/03 11:26:00 EST
874    Tag: leunga-20040903-cygwin-install
875    Description:
876    
877       Added a patch to _arch-n-opsys to enable the Cygwin runtime.
878    The Cygwin runtime is turned on by setting the environment
879    variable SMLNJ_CYGWIN_RUNTIME to 1.
880    
881    ----------------------------------------------------------------------
882    Name: Matthias Blume (blume (at) tti - c (dot) org)
883    Date: 2004/08/31 17:15:00 CDT
884    Tag: blume-20040831-core
885    Description:
886    
887    Added some exports to src/compiler/core.cm upon request by J. Joemann.
888    
889    ----------------------------------------------------------------------
890    Name: Matthias Blume (blume (at) tti - c (dot) org)
891    Date: 2004/08/30 17:55:00 CDT
892    Tag: blume-20040830-installer
893    Description:
894    
895    Upon request by Johannes Joemann:
896    
897    - improved ML code of installer to fall back to coping when renaming
898      fails (i.e., when source and target are on different file systems);
899      the code compiles but has yet to be tested in anger
900    - removed mlrisc from list of default targets (config/targets)
901    
902    ----------------------------------------------------------------------
903    Name: Matthias Blume (blume (at) tti - c (dot) org)
904    Date: 2004/08/27 17:20:00 CDT
905    Tag: blume-20040827-ptreql
906    Description:
907    
908    Added ptreql primop to structure InlineT (upon request from Larry
909    Paulson).
910    
911    ----------------------------------------------------------------------
912    Name: Allen Leung (leunga (at) reservoir (dot) com)
913    Date: 2004/08/15 21:21:00 EST
914    Tag: leunga-110_48-udgraph
915    Description:
916    
917    Another bug fix from Carl Hauser:
918    
919    diff /net/niflab/smlnj48/src/MLRISC/graphs/udgraph.sml udgraph.sml
920    > 48c48
921    > <              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then es else
922    > e::L)
923    > ---
924    > >              | rmv((e as (k,_))::es,L) = rmv(es,if k = i then L else e::L)
925    > Without this, any deletion of an edge in an undirected graph does severe
926    > violence to the graph.
927    
928    ----------------------------------------------------------------------
929    Name: Allen Leung (leunga (at) reservoir (dot) com)
930    Date: 2004/08/10 23:35:00 EST
931    Tag: leunga-110_48-ppc
932    Description:
933    
934       The IBM/MacOS syntax switch on PPC was incorrectly swapped.
935    
936    ----------------------------------------------------------------------
937    Name: Matthias Blume (blume (at) tti - c (dot) org)
938    Date: 2004/08/10 12:00:00 CDT
939    Tag: Release_110_48
940    Description:
941    
942    New working version (110.48).  NEW BOOTFILES!
943    
944    ----------------------------------------------------------------------
945    Name: Allen Leung (leunga (at) reservoir (dot) com)
946    Date: 2004/08/09 12:21:00 EST
947    Tag: leunga-110_47-dijsktra
948    Description:
949    
950    Bug fix from Carl Hauser:
951    
952    single_source_shortest_paths in dijkstra.sml was observed to get wrong
953    answers (by comparing to single_source_shortest_paths in bellman-ford.sml).
954    
955    The problem is that following the expression A.update(dist,s,Num.zero)
956    it is necessary to update the priority queue using Q.decreaseWeight(Q,s).
957    
958    ----------------------------------------------------------------------
959    Name: Matthias Blume (blume (at) tti - c (dot) org)
960    Date: 2004/08/06 18:10:00 CDT
961    Tag: blume-20040806-cmdline
962    Description:
963    
964    Fiddled with handling of command-line options:
965    
966      * sml now quits after processing the command line
967        if -H, -S, -h<n>, or -s<n> appears as the last
968        command-line argument
969      * a new option -q terminates the session when encountered on
970        the command line; subsequent arguments will be ignored
971      * bug fixes: short (erroneous) arguments are no longer ignored
972        completely
973    
974    ----------------------------------------------------------------------
975    Name: Allen Leung (leunga (at) reservoir (dot) com)
976    Date: 2004/08/04 18:17:00 EST
977    Tag: leunga-110_47-ppc-ibm-asm
978    Description:
979    
980      - Added minimal IBM assembly syntax support for PowerPC.
981    
982      - Cygwin: manually changed the file cygwin.def.  Some exported symbols have
983        been altered in the runtime.  We need an automatic way to keep the file
984        in sync.
985    
986    ----------------------------------------------------------------------
987    Name: Matthias Blume (blume (at) tti - c (dot) org)
988    Date: 2004/08/04 14:00:00 CDT
989    Tag: Release_110_47
990    Description:
991    
992    New working version (110.47).  NEW BOOTFILES!
993    
994    ----------------------------------------------------------------------
995    Name: Matthias Blume (blume (at) tti - c (dot) org)
996    Date: 2004/08/03 14:25:00 CDT
997    Tag: blume-20040803-callingconv
998    Description:
999    
1000    Added low-level support for choosing C calling conventions by
1001    twiddling the type of rawccall.  (See
1002    src/compiler/Semant/types/cproto.sml for details.)
1003    
1004    ----------------------------------------------------------------------
1005    Name: Matthias Blume (blume (at) tti - c (dot) org)
1006    Date: 2004/08/02 15:55:00 CDT
1007    Tag: blume-20040802-backout
1008    Description:
1009    
1010    Backed out of change to win32-filesys.c.  The earlier patch to
1011    get_file_time caused CM to produce files with the wrong time stamp.
1012    
1013    ----------------------------------------------------------------------
1014    Name: Matthias Blume (blume (at) tti - c (dot) org)
1015    Date: 2004/08/02 14:45:00 CDT
1016    Tag: blume-20040802-nlffi-win32
1017    Description:
1018    
1019    Added NLFFI support for Win32, adapted from a patch provided by David
1020    Hansel.  This is currently completely untested.  Also, the issue
1021    concerning stdcall vs. ccall is still unresolved.
1022    
1023    ----------------------------------------------------------------------
1024    Name: Matthias Blume (blume (at) tti - c (dot) org)
1025    Date: 2004/07/30 17:55:00 CDT
1026    Tag: blume-20040730-various
1027    Description:
1028    
1029    Gearing up towards 110.47...
1030    
1031    - various minor bugfixes to ml-nlffigen
1032    - a beginning of a manual for nlffi
1033    
1034    - eliminated 'export name=value' in config/install.sh as this does
1035      not work with certain versions of /bin/sh
1036      (Thanks to David King at Motorola for catching this.)
1037    
1038    - several bugfixes provided or suggested by David Hansel at Reactive Systems:
1039      - added a test for tm==NULL to gmtime.c and localtime.c
1040      - applied patch for incorrect GetFileTime under win32
1041      - toSeconds -> toMilliseconds in Win32/win32-process.sml
1042    
1043    ----------------------------------------------------------------------
1044    Name: Matthias Blume (blume (at) tti - c (dot) org)
1045    Date: 2004/07/21 18:20:00 CDT
1046    Tag: blume-20040721-nlffigen
1047    Description:
1048    
1049    - Fixed minor issue in ml-nlffigen: Now generate structure T_foo
1050      for a typedef to an incomplete type, but leave out the "typ" member.
1051      (This is just for consistency.)
1052    - Started to produce what is supposed to become better (i.e., comprehensive)
1053      documentation of what ml-nlffigen does and produces.
1054    
1055    ----------------------------------------------------------------------
1056    Name: Matthias Blume (blume (at) tti - c (dot) org)
1057    Date: 2004/07/14 16:25:00 CDT
1058    Tag: blume-20040714-union
1059    Description:
1060    
1061    Added C_UNION to c-calls/c-types.sml and updated the machinery
1062    (ml-nlffigen, cproto.sml) that conveys C function interface
1063    information to the code generator.
1064    
1065    However, the actual architecture-specific implementation of function
1066    arguments and results that are C unions is still not implemented.
1067    
1068    ----------------------------------------------------------------------
1069    Name: Allen Leung (leunga (at) reservoir (dot) com)
1070    Date: 2004/07/14 14:38:00 EST
1071    Tag: leunga-110_46_1-ppc-lwzu
1072    Description:
1073    
1074        Added these instructions to the PowerPC architecture:
1075        LBZU(X), LHZU(X), LWZU(X),
1076        STWU(X), STFDU, STFSU
1077    
1078        etc...
1079    
1080        Note: I haven't added their instruction encoding into the description.
1081    
1082    ----------------------------------------------------------------------
1083    Name: Allen Leung (leunga (at) reservoir (dot) com)
1084    Date: 2004/07/13 15:04:00 EST
1085    Tag: leunga-110_46_1-ppc-lwarx
1086    Description:
1087    
1088       Added the two instructions LWARX and STWCX to the PowerPC
1089    instruction set.
1090    
1091       A (untested) rewrite of loop-structure.sml.  The old version
1092    is completely broken.
1093    
1094    ----------------------------------------------------------------------
1095    Name: Matthias Blume (blume (at) tti - c (dot) org)
1096    Date: 2004/07/13 13:50:00 CDT
1097    Tag: blume-20040713-nlffi
1098    Description:
1099    
1100    - use paramAlloc to report c-calls with too many arguments
1101      (for PPC version where parameter area is pre-allocated)
1102    - added ccall_maxargspace to machspec (to implement the above)
1103    - made "make" commend in CM's "make" tool configurable
1104    - added option (default: on) for passing the name of the SML/NJ's "bin"
1105      directory to "make"; the call looks like this:
1106    
1107         make <options> SMLNJ_BINDIR=<dir> <target>
1108    
1109      This can be used by the Makefile to, e.g., pick the "right" version
1110      of ml-nlffigen.
1111    - minor code tweaks
1112    
1113    ----------------------------------------------------------------------
1114    Name: Matthias Blume (blume (at) tti - c (dot) org)
1115    Date: 2004/07/12 22:50:00 CDT
1116    Tag: blume-110_46_1-macosx-nlffi
1117    Description:
1118    
1119    NLFFI under Mac OS X now working (sort of).  This is largely untested,
1120    though.
1121    
1122    Note:  1. You have to make a new, clean build of the runtime system.
1123           2. There are new BOOTFILES, you have to use them!
1124              (Doing the bootstrap process yourself would be *very* painful!
1125               If you absolutely have to do it, build the system under
1126               a different architecture and then cross-compile.)
1127    
1128    Version bumped to 110.46.1 to account for runtime data format changes.
1129    
1130    ----------------------------------------------------------------------
1131    Name: Matthias Blume (blume (at) tti - c (dot) org)
1132    Date: 2004/06/18 14:30:00 CDT
1133    Tag: blume-20040618-unix
1134    Description:
1135    
1136    Changed the implementation of structure Unix so that the same stream
1137    is returned every time one of the {text,bin}{In,Out}streamOf functions
1138    is invoked on the same proc.  This is not what the spec currently
1139    says -- although IMO it arguably should.  (See discussion below.)
1140    
1141    ----------------------------------------------------------------------
1142    Name: Matthias Blume (blume (at) tti - c (dot) org)
1143    Date: 2004/06/17 18:15:00 CDT
1144    Tag: Release_110_46
1145    Description:
1146    
1147    New working version (110.46).  NEW BOOTFILES!
1148    
1149    ----------------------------------------------------------------------
1150    Name: Matthias Blume (blume (at) tti - c (dot) org)
1151    Date: 2004/06/17 17:20:00 CDT
1152    Tag: blume-20040617-timer-unix
1153    Description:
1154    
1155    Changed the interface of structures Timer and Unix to match the most
1156    recent Basis spec.
1157    
1158    In the case of Unix there still seems to be an open/weird issue:
1159    
1160       The {text,bin}{In,Out}streamOf functions are supposed to create
1161       fresh streams whenever they are called -- as opposed to have them
1162       return the same stream every time.  This design is supposed to
1163       prevent space leaks caused by proc values hanging on to streams.
1164    
1165       The reap function, on the other hand, is supposed to close the
1166       streams.  This cannot be done without having a handle on the
1167       stream in proc after all...
1168    
1169    I took the liberty to implement the following stopgap solution:
1170    
1171      The proc value hangs on to the most recently created stream(s).
1172      Reap closes those.  If either or both of the two streams hadn't
1173      been created at all yet, then reap will close the corresponding
1174      file descriptors directly.
1175    
1176    PS: I don't understand the original space leak argument anymore.  If
1177    a proc hangs on to the imperative stream, then I/O operations on those
1178    will advance the state of the cached stream and avoid the space leak.
1179    
1180    ----------------------------------------------------------------------
1181    Name: Matthias Blume (blume (at) tti - c (dot) org)
1182    Date: 2004/05/28 16:45:00 CDT
1183    Tag: blume-20040528-basis
1184    Description:
1185    
1186    Added signature PACK_REAL and exported functor PrimIO.
1187    
1188    ----------------------------------------------------------------------
1189    Name: Matthias Blume (blume (at) tti - c (dot) org)
1190    Date: 2004/05/25 16:00:00 CDT
1191    Tag: blume-20040525-group-owner
1192    Description:
1193    
1194    CM now ignores (but still accepts) the "owner" information in group
1195    descriptions.  The owner of a group is its next enclosing
1196    library. Each group must have a unique owner.  (There is a virtual
1197    "toplevel" library that owns groups which are not nested within a real
1198    library.)  Previously, each group had to explicitly declare its owner,
1199    and CM would check that such a declaration is correct.  The new scheme
1200    is to have CM check that for each group there is precisely one owning
1201    library.
1202    
1203    The advantage of the new scheme is that the programmer no longer needs
1204    to maintain the somewhat annoying owner information.  The downside is
1205    that CM cannot enforce the ownership rule across multiple runs of
1206    CM.make.  Fortunately, enclosing the same group in two different
1207    libraries A and B which are not part of the same program does not
1208    cause real problems.
1209    
1210    ----------------------------------------------------------------------
1211    Name: Matthias Blume (blume (at) tti - c (dot) org)
1212    Date: 2004/05/20 16:00:00 CDT
1213    Tag: blume-20040520-win32
1214    Description:
1215    
1216    Made the win32 version work again.  (Strangely, a misplaced comma had
1217    slipped into win32-process.c which prevented the runtime from being
1218    compiled correctly.)
1219    
1220    Also, included a minor addition to ml-build.bat analogous to what was
1221    done in blume-20040519-ml-build.
1222    
1223    ----------------------------------------------------------------------
1224    Name: Matthias Blume (blume (at) tti - c (dot) org)
1225    Date: 2004/05/19 22:10:00 CDT
1226    Tag: blume-20040519-ml-build
1227    Description:
1228    
1229    Arranged for ml-build to clean up after itself a little bit better.
1230    The script generates a temporary SML source file and compiles it using
1231    CM, so CM generates metadata (GUID, SKEL, objectfile) for it.  It now
1232    gets rid of those at the end, so they don't accumulate under .cm.
1233    
1234    This required a minor change to install.sh because the name of the
1235    metadata directory (default: .cm) is actually configurable at
1236    installation time.
1237    
1238    ----------------------------------------------------------------------
1239    Name: Matthias Blume (blume (at) tti - c (dot) org)
1240    Date: 2004/05/18 15:50:00 CDT
1241    Tag: blume-20040518-mkreader
1242    Description:
1243    
1244    Added Posix.IO.mk{Bin,Text}{Reader,Writer} by lifting their respective
1245    implementations from internal modules PosixBinPrimIO and PosixTextPrimIO.
1246    
1247    ----------------------------------------------------------------------
1248    Name: Matthias Blume (blume (at) tti - c (dot) org)
1249    Date: 2004/05/11 14:35:00 CDT
1250    Tag: blume-20040511-win32sock
1251    Description:
1252    
1253    Added previously missing support for many socket-related functions
1254    under win32.  Thanks to David Hansel <hansel@reactive-systems.com>
1255    for the voluminous patch!
1256    
1257    (I have not tested this patch under win32 yet.)
1258    
1259    Here is David's e-mail:
1260    
1261    Hi,
1262    
1263    Attached to this email you find a diff against sml/nj 110.45
1264    that will enable socket support under Windows.
1265    
1266    To apply the patch (using unix or cygwin)
1267    1) gunzip runtime.diff.gz
1268    2) "cd" into "src/runtime" in the source tree of a fresh
1269       110.45 installation.
1270    3) patch -p 1 < [your/path/to]runtime.diff
1271    
1272    The code compiles fine but has NOT yet been extensively tested.
1273    I only ran a few tests for basic socket client functionality
1274    (which worked fine).  Especially the functions that use ioctl
1275    are not tested at all and might not work (see below).
1276    
1277    I implemented this since we want to move to a newer version of sml/nj
1278    but need socket support in order to use it.  This is the first time I
1279    even had a look at the sml/nj source,  so please review my changes
1280    before making this part of the distribution!  Here are a few issues
1281    that I think might be better for someone to solve who is more
1282    familiar with the sml/nj source (and socket programming):
1283    
1284    - getnetbyaddr.c and getnetbyname.c will raise a "not implemented"
1285      exception since I could not figure out what the windows equivalent
1286      of these functions is
1287    
1288    - In sockets-osdep.h there are a some #include statements that are
1289      only used in a few files that include sockets-osdep.h
1290    
1291    - In smlnj-sock-lib.c, function init_fn() calls WSAStartup() but
1292      does not process its return value since I don't know how to
1293      report an error upwards.
1294    
1295    - It would probably be good to have a call to WSACleanup() when
1296      the library is unloaded (if there is such a possibility).
1297      Otherwise I think Windows will take care of this automatically
1298      when the process finishes.
1299    
1300    - I used ioctlsocket() as a replacement for ioctl() but I have
1301      no idea if that is actually the proper replacement on Windows.
1302    
1303    - All these issues are marked in the code by "FIXME" comments.
1304    
1305    We use sml/nj extensively in our products and are quite happy
1306    with it.  I hope this contribution will help you.
1307    
1308    Keep up the good work!
1309    
1310    David
1311    
1312    ----------------------------------------------------------------------
1313    Name: Matthias Blume (blume (at) tti - c (dot) org)
1314    Date: 2004/05/11 14:20:00 CDT
1315    Tag: blume-20040511-installml
1316    Description:
1317    
1318    Fixed two bugs in installml script.  (Thanks to Vesa A. Norrman for
1319    the patch.)
1320    
1321    ----------------------------------------------------------------------
1322    Name: Matthias Blume (blume (at) tti - c (dot) org)
1323    Date: 2004/05/11 14:05:00 CDT
1324    Tag: blume-20040511-nlffi-netbsd
1325    Description:
1326    
1327    Added support for nlffi under netbsd.  (Thanks to Vesa A. Norrman for
1328    the patch.)
1329    
1330    ----------------------------------------------------------------------
1331    Name: Matthias Blume (blume (at) tti - c (dot) org)
1332    Date: 2004/05/11 12:05:00 CDT
1333    Tag: blume-20040511-exports
1334    Description:
1335    
1336    As per request by Adam Chlipala <adam@hcoop.net>, extended various
1337    export lists in compiler-related .cm-files.
1338    
1339    ----------------------------------------------------------------------
1340    Name: Matthias Blume (blume (at) tti - c (dot) org)
1341    Date: 2004/05/11 11:35:00 CDT
1342    Tag: blume-20040511-allsource
1343    Description:
1344    
1345    The installer now honors the "src-smlnj" target again, although its meaning
1346    has changed from "all sources required for the compiler" to "all sources
1347    the installer knows about".  In other words, if you enable "src-smlnj"
1348    in the "targets" file, then the installer will pull in sources for
1349    everything.  (Notice that this refers to source code only.  Compiled
1350    code is still only installed for modules that were requested explicitly
1351    or which are required for other modules that were requested explicitly.)
1352    
1353    ----------------------------------------------------------------------
1354    Name: Matthias Blume (blume (at) tti - c (dot) org)
1355    Date: 2004/04/23 17:40:00 CDT
1356    Tag: blume-20040423-ieee-scan
1357    Description:
1358    
1359    Fixed IEEEReal.scan (and .fromString) so that if there is an overflow
1360    in the exponent calculation we get INF or ZERO (depending on the mantissa
1361    and the sign of the exponent).
1362    
1363    ----------------------------------------------------------------------
1364    Name: Matthias Blume (blume (at) tti - c (dot) org)
1365    Date: 2004/04/23 10:40:00 CDT
1366    Tag: blume-20040423-ml-build
1367    Description:
1368    
1369    The ml-build script now terminates with a non-0 status when something
1370    goes wrong.
1371    
1372    ----------------------------------------------------------------------
1373    Name: Matthias Blume (blume (at) tti - c (dot) org)
1374    Date: 2004/04/22 16:35:00 CDT
1375    Tag: blume-20040422-Option
1376    Description:
1377    
1378    Made exception Option to be the same as exception Option.Option
1379    (as it should be).
1380    
1381    ----------------------------------------------------------------------
1382    Name: Allen Leung (leunga (at) reservoir (dot) com)
1383    Date: 2004/03/19 14:40:00 EST
1384    Tag: leunga-20040319-cygwin-nlffi
1385    Description:
1386    
1387    Fixed the runtime so that ml-nlffi-lib runs on the cygwin version
1388    of SML/NJ.  The problem is that
1389    
1390       lib = dlopen(NULL, ...)
1391       f   = dlsym(lib, "malloc");
1392    
1393    does not work on Windows unless we explicitly export symbols
1394    such as 'malloc' during linking.  We fixed this by explicitly
1395    exporting the required symbols with the magic gcc incantation:
1396    
1397        -Wl,--export-all cygwin.def
1398    
1399    where cygwin.def is a file containing all the symbols that
1400    we wish to export.
1401    
1402    I suspect this is a Windows problem and we'll have to
1403    do the same (somehow with windows compilers) when
1404    we build the native win32 version with the system
1405    calls LoadLibrary/GetProcAddress.
1406    
1407    ----------------------------------------------------------------------
1408    Name: Matthias Blume (blume (at) tti - c (dot) org)
1409    Date: 2004/03/04 16:35:00 CST
1410    Tag: blume-20040304-intinf-fmt
1411    Description:
1412    
1413    Fixed problem with IntInf.fmt (sign would show up on the right instead
1414    of on the left for BIN, OCT, and HEX).
1415    
1416    ----------------------------------------------------------------------
1417    Name: Matthias Blume (blume (at) tti - c (dot) org)
1418    Date: 2004/03/04 11:25:00 CST
1419    Tag: blume-20040304-symlinks
1420    Description:
1421    
1422    Fixed problem with installer script (unix only) where bin/ml-yacc and
1423    friends pointed (via symlinks) to absolute locations instead of just
1424    .run-sml.  This was reported by Vesa A Norrman.
1425    
1426    ----------------------------------------------------------------------
1427    Name: Matthias Blume (blume (at) tti - c (dot) org)
1428    Date: 2004/02/13 14:50:00 CST
1429    Tag: Release_110_45
1430    Description:
1431    
1432    New working version (110.45). New bootfiles.
1433    
1434    ----------------------------------------------------------------------
1435    Name: Matthias Blume (blume (at) tti - c (dot) org)
1436    Date: 2004/01/26 15:15:15 CST
1437    Tag: blume-20040126-toplevel
1438    Description:
1439    
1440    Improved handling of exceptions at the interactive toplevel.
1441    
1442    ----------------------------------------------------------------------
1443    Name: Matthias Blume (blume (at) tti - c (dot) org)
1444    Date: 2004/01/26 11:25:00
1445    Tag: blume-20040126-app
1446    Description:
1447    
1448    Type of top-level "app" corrected.
1449    Added code for setting vp_limitPtrMask to Win32-specific runtime.
1450    
1451    ----------------------------------------------------------------------
1452    Name: Matthias Blume (blume (at) tti - c (dot) org)
1453    Date: 2003/11/18 17:10 CST
1454    Tag: blume-20031118-basis-fiddle
1455    Description:
1456    
1457    - changed Timer interface to what might become the spec
1458    - POSIX_FLAGS -> BIT_FLAGS according to spec
1459    - some other minor discrepancies wrt. spec eliminated
1460    
1461    ----------------------------------------------------------------------
1462    Name: Matthias Blume (blume (at) tti - c (dot) org)
1463    Date: 2003/11/06 12:00:00 CST
1464    Tag: Release_110_44
1465    Description:
1466    
1467    New working version (110.44). New bootfiles.
1468    
1469    ----------------------------------------------------------------------
1470    Name: Matthias Blume (blume (at) tti - c (dot) org)
1471    Date: 2003/11/04 11:50:00 CST
1472    Tag: blume-20031104-move-libraries
1473    Description:
1474    
1475    Eliminated the "dont_move_libraries" directive in config/targets.
1476    (The mechanism was broken and could not be fixed easily.  Moreover,
1477    there does not seem to be any reason not to move all libraries into
1478    lib during installation.  I originally implemented this directive as a
1479    backward-compatibility feature when I first introduced the new CM.
1480    Now that things have been stable for a long time and going back to the
1481    old CM is not an option, there is no reason to keep it around.)
1482    
1483    ----------------------------------------------------------------------
1484    Name: Matthias Blume (blume (at) tti - c (dot) org)
1485    Date: 2003/11/03 16:00:00 CST
1486    Tag: blume-20031103-installdir
1487    Description:
1488    
1489    Made installer honor INSTALLDIR variable again.  (Thanks to Chris
1490    Richards for pointing out the problem and providing the solution.)
1491    
1492    ----------------------------------------------------------------------
1493    Name: Matthias Blume (blume (at) tti - c (dot) org)
1494    Date: 2003/10/01 17:05:00 CDT
1495    Tag: blume-20031001-lal-mlrisc
1496    Description:
1497    
1498    MLRISC bug fix from Lal.
1499    
1500    ----------------------------------------------------------------------
1501    Name: Matthias Blume (blume (at) tti - c (dot) org)
1502    Date: 2003/09/30 16:10:00 CDT
1503    Tag: blume-20030930-primio-bat
1504    Description:
1505    
1506    1. Added openVector, nullRd, and nullWr to PRIM_IO.
1507    2. Improved .bat files (for Win32 port) to make things work under Win95.
1508       (thanks to Aaron S. Hawley for this one)
1509    
1510    ----------------------------------------------------------------------
1511    Name: Matthias Blume (blume (at) tti - c (dot) org)
1512    Date: 2003/09/26 16:05:00 CDT
1513    Tag: blume-20030926-wrappriv
1514    Description:
1515    
1516    Added missing wrapper for privilege "primitive" in $smlnj/viscomp/core.cm.
1517    
1518    ----------------------------------------------------------------------
1519    Name: Matthias Blume (blume (at) tti - c (dot) org)
1520    Date: 2003/09/26 15:00:00 CDT
1521    Tag: blume-20030926-110_43_3
1522    Description:
1523    
1524    - additional cleanup
1525    - version number bump, NEW BOOTFILES
1526    
1527    ----------------------------------------------------------------------
1528    Name: Matthias Blume (blume (at) tti - c (dot) org)
1529    Date: 2003/09/26 12:00:00 CDT
1530    Tag: blume-20030926-ppautoload
1531    Description:
1532    
1533    I modified the read-eval-print loop so that the autoloader gets
1534    invoked whenever the prettyprinter tries to look up a symbol that
1535    is not currently defined in the toplevel environment but which
1536    appears in CM's autoload registry.  As a result, we see far fewer of
1537    those ?.Foo.Bar.xxx names in the prettyprinter's output.
1538    
1539    In addition to this I tried to clean up some pieces of the Basis
1540    implementation (e.g., Socket, Word8Array) in order to prevent other
1541    instances of these ?.Foo.Bar.xxx names from being printed.
1542    
1543    The mechanism that picks names for types still needs some work, though.
1544    (Right now it seems that if there is a type A.t which is defined to
1545    be B.u, but B is unavailable at toplevel, then A.t gets printed as
1546    "?.B.u" although the perhaps more sensible solution would be to use
1547    "A.t" in this case.  In other words, the prettyprinter should follow
1548    a chain of DEFtycs not farther than there are corresponding toplevel
1549    names in the current environment.)
1550    
1551    ----------------------------------------------------------------------
1552    Name: Matthias Blume (blume (at) tti - c (dot) org)
1553    Date: 2003/09/24 16:31:00 CDT
1554    Tag: blume-20030924-installer
1555    Description:
1556    
1557    Another installer tweak:  All the ML code for the installer is now
1558    compiled during CMB.make and put into a little library called
1559    $smlnj/installer.cm.  The installation then simply invokes
1560    
1561       sml -m $smlnj/installer.cm
1562    
1563    and everything happens automagically.
1564    
1565    Win32:  ML code senses value of environment variable SMLNJ_HOME.
1566    Unix:   ML code senses values of environment variables ROOT, CONFIGDIR,
1567            and BINDIR.
1568    
1569    The new scheme guarantees that the ML code responsible for the installation
1570    is in sync with the APIs of the main system.  Also, the installer is
1571    somewhat faster because the installer script is precompiled.
1572    
1573    ----------------------------------------------------------------------
1574    Name: Matthias Blume (blume (at) tti - c (dot) org)
1575    Date: 2003/09/24 15:35:00 CDT
1576    Tag: blume-20030924-synsock
1577    Description:
1578    
1579    Added a signature SYNCHRONOUS_SOCKET to basis.cm.  This is like SOCKET
1580    but excludes all non-blocking operations.  Defined SOCKET (in Basis)
1581    and CML_SOCKET in terms of SYNCHRONOUS_SOCKET.  Removed superfluous
1582    implementations of non-blocking operations from CML's Socket
1583    structure.
1584    
1585    ----------------------------------------------------------------------
1586    Name: Matthias Blume (blume (at) tti - c (dot) org)
1587    Date: 2003/09/24 15:10:05 CDT
1588    Tag: blume-20030924-sockets
1589    Description:
1590    
1591    1. Fixed SOCKET API and implementation to match Basis spec.
1592       This required changing the internal representation of sockets to one
1593       that remembers (for each socket file descriptor) whether it is currently
1594       blocking or non-blocking.  This state is maintained lazily (i.e., a system
1595       call is made only if the state actually needs to change).
1596    
1597    2. OS-specific details of sockets were moved into separate files, thus
1598       making it possible to unify the bulk of the socket implementations
1599       between Unix and Win32.
1600    
1601    3. CML's socket API changed accordingly.
1602       (Note that we need to remove non-blocking functions from this API
1603       since they are redundant in the case of CML!)
1604    
1605    4. CML's socket implementation now makes use of non-blocking functions
1606       provided by Basis, thus removing all OS-dependent code from this part
1607       of CML.
1608    
1609    5. Changed Real64.precision from 52 to 53.  Minor cleanup in Real64 code.
1610    
1611    ----------------------------------------------------------------------
1612    Name: Matthias Blume (blume (at) tti - c (dot) org)
1613    Date: 2003/09/22 12:10:00 CDT
1614    Tag: blume-20030922-110_43_2
1615    Description:
1616    
1617    Made a new interim version and bootfiles for developer's bootstrapping
1618    convenience.
1619    
1620       110.43.2 -- NEW BOOTFILES
1621    
1622    ----------------------------------------------------------------------
1623    Name: Matthias Blume (blume (at) tti - c (dot) org)
1624    Date: 2003/09/19 15:55:00 CDT
1625    Tag: blume-20030919-cmdir
1626    Description:
1627    
1628    1. new-install.sh -> install.sh
1629    2. changed default CM "metadata" directory name to ".cm" (instead of "CM")
1630    3. tweaked installer so that another name instead of .cm can be chosen
1631       at install time (by setting the CM_DIR_ARC environment variable
1632       during installation); once installation is complete, the name is
1633       fixed
1634    
1635    ----------------------------------------------------------------------
1636    Name: Matthias Blume (blume (at) tti - c (dot) org)
1637    Date: 2003/09/18 16:00:00 CDT
1638    Tag: blume-20030918-110_43_1
1639    Description:
1640    
1641    Made a new interim version and bootfiles for developer's bootstrapping
1642    convenience.
1643    
1644       110.43.1 -- NEW BOOTFILES
1645    
1646    ----------------------------------------------------------------------
1647    Name: Matthias Blume (blume (at) tti - c (dot) org)
1648    Date: 2003/09/18 15:20:00 CDT
1649    Tag: blume-20030918-misc
1650    Description:
1651    
1652    1. Exported fractionsPerSecond etc. from TimeImp (but not from Time as
1653       this seems to be controversial at the moment) and used those in
1654       Posix.ProcEnv.times.
1655    
1656    2. Added Time.{from,to}Nanoseconds to Time.
1657    
1658    3. Improved Real.{from,to}LargeInt by avoiding needless calculations.
1659       For example, fromLargeInt never needs to look at more than 3 "big
1660       digits" to get its 53 bits of precision.
1661    
1662    ----------------------------------------------------------------------
1663    Name: Matthias Blume (blume (at) tti - c (dot) org)
1664    Date: 2003/09/17 16:30:00 CDT
1665    Tag: blume-20030917-real32-slices
1666    Description:
1667    
1668    Added an entry to the primitive environment
1669    (compiler/Semant/statenv/prim.sml) for int32->real64 conversion and
1670    added code to compiler/CodeGen/main/mlriscGen.sml to implement it.
1671    
1672    Removed some of the "magic" constants in real64.sml and replaced them
1673    with code that generates these values from their corresponding
1674    integer counterparts.
1675    
1676    Made all(?) the slice-related changes to the Basis and made everything
1677    compile again...
1678    
1679    ----------------------------------------------------------------------
1680    Name: Matthias Blume (blume (at) tti - c (dot) org)
1681    Date: 2003/09/15 17:45:00 CDT
1682    Tag: blume-20030915-rbase
1683    Description:
1684    
1685    Fixed bug in Real.fromLargeInt.
1686    
1687    ----------------------------------------------------------------------
1688    Name: Matthias Blume (blume (at) tti - c (dot) org)
1689    Date: 2003/09/13 18:11:00 CDT
1690    Tag: blume-20030913-libinstall
1691    Description:
1692    
1693    Minor bugfix in config/libinstall (set anchor with path to
1694    standalone tool after installing it, otherwise libraries that
1695    need ml-lex or ml-yacc won't compile the first time the installer
1696    runs).
1697    
1698    ----------------------------------------------------------------------
1699    Name: Matthias Blume (blume (at) tti - c (dot) org)
1700    Date: 2003/09/12 11:45:00 CDT
1701    Tag: blume-20030912-various
1702    Description:
1703    
1704     - fixed bug in Real.toLargeInt
1705     - fixed bug in Posix.ProcEnv.times
1706     - changed inputLine functions to return an option
1707     - minor installer improvements / bugfixes
1708     - changed default @SMLalloc parameter for x86/celeron to 64k
1709    
1710    ----------------------------------------------------------------------
1711    Name: Matthias Blume (blume (at) tti - c (dot) org)
1712    Date: 2003/09/09 22:00:00 CDT
1713    Tag: Release_110_43
1714    Description:
1715    
1716        New working release 110.43.  New bootfiles.
1717    
1718    ----------------------------------------------------------------------
1719    Name: Matthias Blume (blume (at) tti - c (dot) org)
1720    Date: 2003/09/09 19:20:00 CDT
1721    Tag: blume-20030909-installer
1722    Description:
1723    
1724    Rewrote large parts of config/install.sh in SML (config/libinstall.sml).
1725    Modified config/install.bat to take advantage of it.  Also modified
1726    config/install.sh (and called it config/new-install.sh) to take advantage
1727    of it on Unix systems.  (The SML code is (supposed to be) platform-
1728    independent.)
1729    
1730    The installer can now install everything under Win32
1731    as well as under *nix as long as it compiles.
1732    
1733    Other changes:
1734    
1735      - made CML compile again under Win32
1736      - made eXene compile under Win32 (by providing a fake structure UnixSock
1737        and by using OS.Process.getEnv instead of Posix.ProcEnv.getenv)
1738      - fixed a bug in nowhere: it assumed that type OS.Process.status is the
1739        same as type int;  under Win32 it isn't
1740      - fixed some slice-related problems in the win32-specific parts of CML
1741      - added a functor argument "sameVol" to os-path-fn.sml in the Basis
1742        (under Win32, the volume name is case-insensitive, and the
1743        OS.Path code compares volume names for equality)
1744    
1745    ----------------------------------------------------------------------
1746    Name: Matthias Blume (blume (at) tti - c (dot) org)
1747    Date: 2003/09/08 11:55:00 CDT
1748    Tag: blume-20030908-fullpath
1749    Description:
1750    
1751    Made Win32 version of OS.FileSys.fullPath return current directory
1752    when given an empty string.  This is what the spec says, and incidentally,
1753    CM depends on it.  (CM otherwise goes into an infinite loop in certain
1754    cases when presented with the name of a non-existing .cm file.)
1755    
1756    ----------------------------------------------------------------------
1757    Name: Matthias Blume (blume (at) tti - c (dot) org)
1758    Date: 2003/09/04 16:30:00 CDT
1759    Tag: blume-20030905-slices-etc
1760    Description:
1761    
1762      1. Changed interface to vectors and arrays in Basis to match
1763         (draft) Basis spec.
1764      2. Added signatures and implementations of slices according to
1765         Basis spec.
1766      3. Edited source code throughout the system to make it compile again
1767         under 1. and 2.  (In some cases code had to be added to have it
1768         match the new signatures.)
1769      4. MLRISC should be backward-compatible: the copies of the originals
1770         of files that needed to change under 3. were retained, the .cm files
1771         check the compiler version number and use old versions when
1772         appropriate.
1773      5. Changed type of OS.FileSys.readDir and Posix.FileSys.readdir to
1774         dirstream -> string option (in accordance with Basis spec).
1775      6. When generating code that counts lines, ml-lex used function
1776         CharVector.foldli, taking advantage of its old interface.
1777         This has been replaced with the corresponding code from
1778         CharVectorSlice.  (html-lex must be re-lexed!)
1779      7. BitArray in smlnj-lib/Util has been extended/modified to match the
1780         new MONO_ARRAY signature.  (Do we need BitArraySlice?)
1781      8. Removed temporary additions (fromInternal, toInternal) from the
1782         (now obsolete) IntInf in smlnj-lib/Util.
1783      9. Cleaned up structure Byte.
1784     10. Added localOffset, scan, and fromString to Date (according to spec).
1785         Cleaned/corrected implementation of Date.
1786         (Still need to check for correctness; implement better canonicalizeDate.)
1787     11. Added "scan" to signature IEEE_REAL.
1788     12. Some improvements to IntInf [in particular: efficiency-hack for
1789         mod and rem when second operand is 2 (for parity checks).]
1790     13. Changed representation of type Time.time, using a single IntInf.int
1791         value counting microseconds.  This considerably simplified the
1792         implementation of structure Time.  We now support negative time
1793         values; scan and fromString handle signs.
1794     14. Functor PrimIO now takes two additional arguments (VectorSlice and
1795         ArraySlice).
1796    
1797    ----------------------------------------------------------------------
1798    Name: Matthias Blume (blume (at) tti - c (dot) org)
1799    Date: 2003/08/28 17:00:00 CDT
1800    Tag: blume-20030828-intinf
1801    Description:
1802    
1803    This is a major update which comes with a version number bump
1804    (110.42.99 -- yes, we are really close to 110.43 :-), NEW BOOTFILES,
1805    and an implementation of IntInf in the Basis.
1806    
1807    There are a fairly large number of related changes and updates throughout
1808    the system:
1809    
1810    Basis:
1811            - Implemented IntInf.
1812            - Made LargeInt a projection of IntInf (by filtering through INTEGER).
1813            - Added some missing Real64 operations, most notably Real.toLargeInt.
1814            - Added FixedInt as a synonym for Int32.
1815    
1816    compiler:
1817            * Added support for a built-in intinf type.
1818              - literals
1819              - pattern matching
1820              - conversion shortcuts (Int32.fromLarge o Int.toLarge etc.)
1821              - overloading on literals and operations
1822    
1823            This required adding a primitive type intinf, some additional
1824            primops, and implementations for several non-trivial intinf
1825            operations in Core.  (The intinf type is completely abstract
1826            to the compiler; all operations get delegated back to the Core.)
1827    
1828            * Intinf equality is handled by polyequal.  However, the compiler
1829            does not print its usual warning in this case (since polyequal
1830            is the right thing to do there).
1831    
1832            * Improved the organization of structure InlineT.
1833    
1834            * A word about conversion primops:
1835            If conversions involving intinf do not cancel out during
1836            CPS contract, then the compiler must insert calls to Core functions.
1837            Since all core access must be resolved already during the FLINT
1838            translate phase, it would be too late a the time of CPS contract
1839            to add new Core calls.  For this reason, conversion primops
1840            for intinf carry two arguments:  1. the numeric argument that
1841            they are supposed to convert, and 2. the Core function that
1842            can help with this conversion if necessary.  If CPS contract
1843            eliminates a primop, then the associated Core function becomes
1844            dead and goes away.  Intinf conversion primops that do not get
1845            eliminated by CPS contract get rewritten into calls of their
1846            core functions by a separate, new phase.
1847    
1848    interactive system:
1849            - Control.Print.intinfDepth controls max length of intinf constants
1850            being printed. (Analogous to Control.Print.stringDepth.)
1851            - Cleanup in printutil and pputil: got rid of unused stuff and
1852            duplicates; replaced some of the code with code that makes better
1853            use of library functionality.
1854    
1855    CM:
1856            Bugfix: parse-errors in init group (system/smlnj/init/init.cmi)
1857            are no longer silent.
1858    
1859    CKIT:
1860            Fixed mismatched uses of Int32 and LargeInt.  I always decided
1861            in favor of LargeInt -- which is now the same as IntInf.
1862            CKIT-knowledgable people should check whether this is what's
1863            intended and otherwise change things back to using Int32 or
1864            FixedInt.
1865    
1866    Throughout the code:
1867            Started using IntInf.int literals and built-in operations
1868            (e.g., comparison with 0) where this seems appropriate.
1869    
1870    
1871    ----------------------------------------------------------------------
1872    Name: Dave MacQueen (dbm@cs.uchicago.edu)
1873    Date: 2003/08/13 11:36:00 CDT
1874    Tag: dbm-20030813-mcz-merge1
1875    Description:
1876    
1877    Merging changes from the mcz-branch development branch into trunk.
1878    These changes involve replacement of the emulated old prettyprinter
1879    interface with direct use of the SML/NJ Lib PP library, and fixing
1880    of a couple of bugs (895, 1186) relating to error messages.  A new
1881    prettyprinter for ast datatypes (Elaborator/print/ppast.{sig,sml})
1882    has been added.
1883    
1884    ----------------------------------------------------------------------
1885    Name: Matthias Blume (blume (at) tti - c (dot) org)
1886    Date: 2003/08/11 15:45:00 CDT
1887    Tag: blume-20030811-windows
1888    Description:
1889    
1890    Version number bumped to 110.42.9.  NEW BOOTFILES!!!
1891    
1892        http://smlnj.cs.uchicago.edu/dist/working/110.42.9/
1893    
1894    This patch restores SML/NJ's ability to run under win32.  There are a
1895    number of changes, including fixes for several bugs that had gone
1896    unnoticed until now:
1897    
1898      - uname "CYGWIN_NT*" is recognized as win32 (This is relevant only when
1899        trying to run the win32 version from within cygwin.)
1900    
1901      - There are a number of simple .bat scripts that substitute for their
1902        corresponding Unix shell-scripts.  (See below.)
1903    
1904      - The internals of ml-build have been modified slightly.  The main
1905        difference is that instead of calling ".link-sml" (or link-sml.bat)
1906        using OS.Process.system, the ML process delegates this task back
1907        to the script.  Otherwise problems arise in mixed environments such
1908        as Cygwin where scripts look and work like Unix scripts, but
1909        where OS.Process.system cannot run them.
1910    
1911      - In CM, the srcpath pickler used native pathname syntax -- which
1912        is incorrect in the case of cross-compilation.  The new pickle format
1913        is independent of platform-specific naming conventions.
1914    
1915      - Path configuration files (such as lib/pathconfig) can now choose
1916        between native and standard syntax.  Placing a line of the form
1917    
1918            standard!
1919    
1920        into the file causes all subsequent paths to be interpreted using
1921        CM standard pathname syntax (= Unix conventions); a line
1922    
1923            native!
1924    
1925        switches back to native style.  This was needed so that
1926        path config files can be written portably, see src/system/pathconfig.
1927    
1928      - Runtime system:
1929    
1930          - win32-filesys.c:  get_file_time and set_file_time now
1931            access modification time, not creation time.
1932    
1933          - I/O code made aware of new array representation.
1934    
1935          - Bug fixes in X86.prim.masm.
1936    
1937       - src/system/makeml made aware of win32. (For use under cygwin
1938            and other Unix-environments for windows.)
1939    
1940       - In Basis, fixed off-by-one error in win32-io.sml (function vecF)
1941         which caused BinIO.inputAll to fail consistently.
1942    
1943    .bat scripts:
1944    
1945       Windows .bat scripts assume that SMLNJ_HOME is defined.
1946    
1947       - sml.bat, ml-yacc.bat, ml-lex.bat:  Driver scripts for standalone
1948         applications (sml, ml-yacc, ml-lex).
1949       - ml-build.bat: analogous to ml-build.
1950       - config\install.bat:  Analogous to config/install.sh.  This requires
1951         that SMLNJ_HOME is set and that Microsoft Visual C is ready to use.
1952         (nmake etc. must be on the path, and vcvars32 must have been run.)
1953         Moreover, sources for ml-lex and ml-yacc need to exist under src,
1954         and the bootfile hierarchy must have been unpacked under
1955         sml.boot.x86-win32.
1956         The script is very primitive and does a poor job at error checking.
1957         It only installs the base system, ml-lex, and ml-yacc.  No other
1958         libraries are being installed (i.e., you get only those that
1959         are part of the compiler.)
1960       - link-sml.bat:  analogous to .link-sml, but not currently used
1961    
1962    Unrelated bug fixes:
1963    
1964      - ml-nlffigen now exports structures ST_* corresponding to incomplete
1965        types.
1966      - Added getDevice to PP/src/pp-debug-fn.sml.  (Would not compile
1967        otherwise.)
1968    
1969    ----------------------------------------------------------------------
1970    Name: Dave MacQueen (macqueen@cs.uchicago.edu)
1971    Date: 2003/06/17
1972    Tag: macqueen-20030617-bug895
1973    Description:
1974    
1975    Modified compiler/Elaborator/print/pptype.sml to fix bug 895.
1976    Tag will be used for new development branch (mcz-branch) for
1977    use by MacQueen, (Lucasz) Zairek, and (George) Cao at uchicago.
1978    
1979    ----------------------------------------------------------------------
1980    Name: Matthias Blume (blume (at) tti - c (dot) org)
1981    Date: 2003/05/27 16:55:00 CDT
1982    Tag: blume-20030527-polyeq
1983    Description:
1984    
1985    Tried to eliminated most cases of polymorphic equality.
1986    
1987    ----------------------------------------------------------------------
1988    Name: Matthias Blume (blume (at) tti - c (dot) org)
1989    Date: 2003/05/21 17:45:00 CDT
1990    Tag: blume-20030517-complete
1991    Description:
1992    
1993    Two changes:
1994    
1995      1. Added a flag for controlling whether non-exhaustive bindings will
1996         be treated as errors (default is false).
1997      2. Cleaned up the *entire* source tree so that CMB.make goes through
1998         without a single non-exhaustive match- or bind warning.
1999    
2000    ----------------------------------------------------------------------
2001    Name: Matthias Blume (blume (at) tti - c (dot) org)
2002    Date: 2003/05/17 10:20:00 CDT
2003    Tag: blume-20030517-absyn
2004    Description:
2005    
2006    1. Added cases for IF, WHILE, ANDALSO, and ORELSE to Absyn.
2007    
2008       This mainly affects the quality of error messages.  However, some
2009       of the code is now more straightforward than before. (Treatment of
2010       the above four constructs in translate.sml is much simpler than
2011       the "macro-expansion" that was going on before.  Plus, the mach-
2012       compiler no longer gets invoked just to be able to compile an
2013       if-expression.)
2014    
2015    2. The ErrorMsg.Error exception is now caught and absorbed by the
2016       interactive loop.
2017    
2018    ----------------------------------------------------------------------
2019    Name: Allen Leung
2020    Date: 2003/05/16 13:05:00 CDT
2021    Tag: leunga-20030516-cygwin-runtime
2022    Description:
2023    
2024        Ported the runtime system to cygwin, which uses the unix
2025    x86-unix bin files.  Missing/buggy features:
2026    
2027       o getnetbyname, getnetbyaddr: these functions seem to be missing in
2028         the Cygwin library.
2029       o Ctrl-C handling may be flaky.
2030       o Windows system calls and Windows I/O are not supported.
2031    
2032       A new set of binfiles is located at:
2033    
2034         http://www.dorsai.org/~leunga/boot.x86-unix.tgz
2035    
2036       This is only needed for bootstrapping the cygwin version of smlnj.
2037       Other x86 versions can use the existing binfiles.
2038    
2039    ----------------------------------------------------------------------
2040    Name: Matthias Blume
2041    Date: 2003/04/08 15:42:00 CDT
2042    Tag: blume-20030408-listpair
2043    Description:
2044    
2045    1. Added a target 'mlrisc' to installer.
2046    
2047    2. Added missing elements to structure ListPair.
2048    
2049    ----------------------------------------------------------------------
2050    Name: Allen Leung
2051    Date: 2003/01/07 10:40:00 EST
2052    Tag: leunga-20030107-int-rem
2053    Description:
2054    
2055    Fixed a bug in Int.rem(x,y) where y is a power of 2 on x86.
2056    The arguments to the SUBL instruction were swapped.
2057    
2058    ----------------------------------------------------------------------
2059    Name: Matthias Blume
2060    Date: 2002/12/12 16:25:00 EST
2061    Tag: blume-20021212-risc-ra
2062    Description:
2063    
2064    Fixed a serious bug in the rewrite code for FP spilling/reloading that
2065    sent the RA into an infinite loop when floating point registers get
2066    spilled.  (Because of this bug, e.g., nucleic stopped compiling between
2067    110.37 and 110.38.)
2068    There was another set of potential problems related to the handling of
2069    MLRISC annotations (but those did not yet cause real problems, apparently).
2070    
2071    ----------------------------------------------------------------------
2072    Name: Matthias Blume
2073    Date: 2002/12/06 22:40:00 EST
2074    Tag: blume-20021206-cm-fileid
2075    Description:
2076    
2077    Added a call of SrcPath.sync at the beginning of Parse.parse (in CM).
2078    This fixes the problem of CM getting confused by files that suddenly
2079    change their identity (e.g., by getting unlinked and recreated by some
2080    text editor such as vi).  There might be a better/cheaper/cleaner way
2081    of doing this, but for now this will have to do.
2082    
2083    ----------------------------------------------------------------------
2084    Name: Matthias Blume
2085    Date: 2002/10/28 09:50:00 EST
2086    Tag: blume-20021028-typecheck
2087    Description:
2088    
2089    Exported structure Typecheck from $smlnj/viscomp/core.cm.
2090    
2091    ----------------------------------------------------------------------
2092    Name: Matthias Blume
2093    Date: 2002/10/17 09:10:00 EDT
2094    Tag: Release_110_42
2095    Description:
2096    
2097    In good old tradition, there has been a slight hiccup so that we have
2098    to patch 110.42 after the fact.   The old release tag has been replaced
2099    (see below).
2100    
2101    The change solves a problem with two competing approaches the
2102    configuration problem regarding MacOS 10.1 vs. MacOS 10.2 which got in
2103    each other's way.
2104    
2105    This change only affects the runtime system code and the installer script.
2106    (No new bootfiles.)
2107    
2108    ----------------------------------------------------------------------
2109    Name: Matthias Blume
2110    Date: 2002/10/16 12:00:00 EDT
2111    Tag: Release_110_42_removed
2112    Description:
2113    
2114    New working release.  New bootfiles.
2115    
2116    ----------------------------------------------------------------------
2117    Name: Matthias Blume
2118    Date: 2002/10/10 13:10:00 EDT
2119    Tag: blume-20021010-ppc-divs
2120    Description:
2121    
2122    The mltree operator DIVS must be implemented with an overflow check on
2123    the PPC because the hardware indicates divide-by-zero using "overflow" as
2124    well.
2125    
2126    ----------------------------------------------------------------------
2127    Name: Matthias Blume
2128    Date: 2002/07/23 11:45:00
2129    Tag: blume-20020723-smlnj-home
2130    Description:
2131    
2132    Sml now senses the SMLNJ_HOME environment variable.  If this is set,
2133    then the bin dir is assumed to be in $SMLNJ_HOME/bin and (unless
2134    CM_PATHCONFIG is also set), the path configuration file is assumed
2135    to be in $SMLNJ_HOME/lib/pathconfig.  This way one can easily move
2136    the entire tree to some other place and everything will "just work".
2137    
2138    (Companion commands such as ml-build and ml-makedepend also sense this
2139    variable.)
2140    
2141    ----------------------------------------------------------------------
2142    Name: Matthias Blume
2143    Date: 2002/07/12 21:19:00 EDT
2144    Tag: blume-20020712-liveness
2145    Description:
2146    
2147    Exported two useful "step" functions from liveness module (MLRISC).
2148    
2149    ----------------------------------------------------------------------
2150    Name: Matthias Blume
2151    Date: 2002/07/05 16:00 EDT
2152    Tag: Release_110_41
2153    Description:
2154    
2155    New working release.  New bootfiles.
2156    
2157    ----------------------------------------------------------------------
2158    Name: Matthias Blume
2159    Date: 2002/07/05 10:25:00 EDT
2160    Tag: blume-20020705-btimp
2161    Description:
2162    
2163    Exported structure BTImp from $smlnj/viscomp/debugprof.cm so that
2164    other clients can set up backtracing support.
2165    
2166    ----------------------------------------------------------------------
2167    Name: Matthias Blume
2168    Date: 2002/06/25 17:23:00 EDT
2169    Tag: blume-20020625-fpmax
2170    Description:
2171    
2172    Fixed a bug in translation of INLMAX (and INLMIN) for the floating-point
2173    case.  (The sense of the isNaN test was reversed -- which made min and
2174    max always return their first argument.)
2175    
2176    ----------------------------------------------------------------------
2177    Name: Matthias Blume
2178    Date: 2002/06/11
2179    Tag: blume-20020611-unixpath
2180    Description:
2181    
2182    Back-ported OS.Path.{from,to}UnixPath from idlbasis-devel branch.
2183    
2184    ----------------------------------------------------------------------
2185    Name: Matthias Blume
2186    Date: 2002/06/10 16:35:00 EDT
2187    Tag: blume-20020610-ieeereal
2188    Description:
2189    
2190    I back-ported my implementation of IEEEReal.fromString from the
2191    idlbasis-devel branch so that we can test it.
2192    
2193    Another small change is that ppDec tries to give more information
2194    than just "<sig>" in the case of functors.  However, this code is
2195    broken in some mysterious way if the functor's body's signature
2196    has not been declared by ascription but gets inferred from the
2197    implementation.  This needs fixing...
2198    
2199    ----------------------------------------------------------------------
2200    Name: Matthias Blume
2201    Date: 2002/05/31
2202    Tag: blume-20020531-btrace-mode
2203    Description:
2204    
2205    Resurrected SMLofNJ.Internals.BTrace.mode.  (It accidentally fell by
2206    the wayside when I switched over to using Controls everywhere.)
2207    
2208    ----------------------------------------------------------------------
2209    Name: Lal George
2210    Date: 2002/05/23 12:21:40 EDT
2211    Tag: george-20020523-visual-labels
2212    Description:
2213    
2214    Labels are now displayed in the graphical output to make
2215    the fall-through and target blocks obvious.
2216    
2217    ----------------------------------------------------------------------
2218    Name: Matthias Blume
2219    Date: 2002/05/22 11:03:00 EDT
2220    Tag: blume-20020522-shrink
2221    Description:
2222    
2223    John tweaked yesterday's fix for 1131 to handle an out-of-memory
2224    situation that comes up when allocating huge arrays.
2225    
2226    ----------------------------------------------------------------------
2227    Name: Matthias Blume
2228    Date: 2002/05/21 16:00:00 EDT
2229    Tag: Release_110_40
2230    Description:
2231    
2232    New working release (110.40).  New bootfiles.
2233    
2234    [Also: John Reppy fixed GC bug 1131.]
2235    
2236    ----------------------------------------------------------------------
2237    Name: Matthias Blume
2238    Date: 2002/05/21 12:35:00 EDT
2239    Tag: blume-20020521-cmdoc
2240    Description:
2241    
2242    CM documentation update.
2243    
2244    ----------------------------------------------------------------------
2245    Name: Matthias Blume
2246    Date: 2002/05/21 10:55:00 EDT
2247    Tag: blume-20020521-misc
2248    Description:
2249    
2250    - John tweaked runtime to be silent on heap export (except when
2251      GC messages are on).
2252    - I added a few more things (cross-compiling versions of CMB) to
2253      config/preloads (as suggestions).
2254    
2255    ----------------------------------------------------------------------
2256    Name: Matthias Blume
2257    Date: 2002/05/20 22:25:00 EDT
2258    Tag: blume-20020520-controls
2259    Description:
2260    
2261    - Added ControlUtil structure to control-lib.cm.
2262    - Use it throughout.
2263    - Used Controls facility to define MLRISC controls (as opposed to
2264      registering MLRISC control ref cells with Controls after the fact)
2265    - Fixed messed-up controls priorities.
2266    
2267    * Removed again all the stuff from config/preloads that one wouldn't
2268      be able to preload at the time the initial heap image is built.
2269      (Many libraries, e.g., CML, do not exist yet at this time.  The
2270       only libraries that can be preloaded via config/preloads are those
2271       that come bundled with the bootfiles.)
2272    
2273    ----------------------------------------------------------------------
2274    Name: Matthias Blume
2275    Date: 2002/05/20 10:59:00 EDT
2276    Tag: blume-20020520-preloads
2277    Description:
2278    
2279    Added a lot of commented-out suggestions for things to be included
2280    in config/preloads.
2281    
2282    ----------------------------------------------------------------------
2283    Name: Allen Leung
2284    Date: 2002/05/18 14:20:00 EDT
2285    Tag: leunga-20020518-mdl
2286    Description:
2287    
2288    o Made the mdl tool stuff compile and run again.
2289    o I've disabled all the stuff that depends on RTL specifications; they
2290      are all badly broken anyway.
2291    
2292    ----------------------------------------------------------------------
2293    Name: Matthias Blume
2294    Date: 2002/05/17 16:49:00 EDT
2295    Tag: blume-20020517-controls
2296    Description:
2297    
2298    0. John Reppy made several modifications to the SML/NJ library.
2299       In particular, there is a shiny new controls-lib.cm.
2300    
2301    1. Pushed new controls interface through compiler so that everything
2302       compiles again.
2303    
2304    2. Added FormatComb and FORMAT_COMB to the CML version of the
2305       SML/NJ library (so that CML compiles again).
2306    
2307    3. Modified init scripts because XXX_DEFAULT environment variables
2308       are no longer with us.  (Boot-time initialization is now done
2309       using the same environment variables that are also used for
2310       startup-time initialization of controls.)
2311    
2312    ----------------------------------------------------------------------
2313    Name: Lal George
2314    Date: 2002/05/15 09:20:10 EDT
2315    Tag: george-20020515-pseudo-op-decls
2316    Description:
2317    
2318            All pseudo-ops emitted before the first segment declaration
2319    such as TEXT, DATA, and BSS directives are assumed to be global
2320    declarations and are emitted first in the assembly file. This is
2321    useful in a number of situations where one has pseudo-ops that are not
2322    specific to any segment, and also works around the constraint that one
2323    cannot have client pseudo-ops in the TEXT segment.
2324    
2325            Because no segment is associated with these declarations it is
2326    an error to allocate any space or objects before the first segment
2327    directive and an exception will be raised. However, we cannot make
2328    this check for client pseudo-ops.
2329    
2330            These top level declarations are a field in the CFG graph_info.
2331    In theory you can continue to add to this field after the CFG has been
2332    built -- provided you know what you are doing;-)
2333    
2334    ----------------------------------------------------------------------
2335    Name: Matthias Blume
2336    Date: 2002/05/13 16:40:00 EDT
2337    Tag: blume-20020513-pp-etc
2338    Description:
2339    
2340    A few minor bugfixes:
2341    
2342      - Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
2343        (Bogus printouts for redefined bindings still occur.  Compiler
2344        bug should no longer occur now.  We need to redo the prettyprinter
2345        from scratch.)
2346    
2347      - CM pathname printer now also adds escape sequences for ( and )
2348    
2349      - commend and docu fixes for ml-nlffi
2350    
2351    ----------------------------------------------------------------------
2352    Name: Matthias Blume
2353    Date: 2002/05/10 16:40:00 EDT
2354    Tag: blume-20020510-erg-textio
2355    Description:
2356    
2357    Applied the following bugfix provided by Emden Gansner:
2358    
2359        Output is corrupted when outputSubstr is used rather than output.
2360    
2361        The problem occurs when a substring
2362    
2363            ss = (s, dataStart, dataLen)
2364    
2365        where dataStart > 0, fills a stream buffer with avail bytes left.
2366        avail bytes of s, starting at index dataStart, are copied into the
2367        buffer, the buffer is flushed, and then the remaining dataLen-avail
2368        bytes of ss are copied into the beginning of the buffer. Instead of
2369        starting this copy at index dataStart+avail in s, the current code
2370        starts the copy at index avail.
2371    
2372        Fix:
2373        In text-io-fn.sml, change line 695 from
2374             val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
2375        to
2376             val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
2377    
2378    ----------------------------------------------------------------------
2379    Name: Matthias Blume
2380    Date: 2002/04/12 13:55:00 EDT
2381    Tag: blume-20020412-assyntax
2382    Description:
2383    
2384    1. Grabbed newer assyntax.h from the XFree86 project.
2385    2. Fiddled with how to compile X86.prim.asm without warnings.
2386    3. (Very) Minor cleanup in CM.
2387    
2388    ----------------------------------------------------------------------
2389    Name: Matthias Blume
2390    Date: 2002/04/01 (no joke!) 17:07:00 EST
2391    Tag: blume-20020401-x86div
2392    Description:
2393    
2394    Added full support for div/mod/rem/quot on the x86, using the machine
2395    instruction's two results (without clumsily recomputing the remainder)
2396    directly where appropriate.
2397    
2398    Some more extensive power-of-two support was added to the x86 instruction
2399    selector (avoiding expensive divs, mods, and muls where they can be
2400    replaced with cheaper shifts and masks).  However, this sort of thing
2401    ought to be done earlier, e.g., within the CPS optimizer so that
2402    all architectures benefit from it.
2403    
2404    The compiler compiles to a fixed point, but changes might be somewhat
2405    fragile nevertheless.  Please, report any strange things that you might
2406    see wrt. div/mod/quot/rem...
2407    
2408    ----------------------------------------------------------------------
2409    Name: Matthias Blume
2410    Date: 2002/03/29 17:22:00
2411    Tag: blume-20020329-div
2412    Description:
2413    
2414    Fixed my broken div/mod logic.  Unfortunately, this means that the
2415    inline code for div/mod now has one more comparison than before.
2416    Fast paths (quotient > 0 or remainder = 0) are not affected, though.
2417    The problem was with quotient = 0, because that alone does not tell
2418    us which way the rounding went.  One then has to look at whether
2419    remainder and divisor have the same sign...  :(
2420    
2421    Anyway, I replaced the bootfiles with fresh ones...
2422    
2423    ----------------------------------------------------------------------
2424    Name: Matthias Blume
2425    Date: 2002/03/29 14:10:00 EST
2426    Tag: blume-20020329-inlprims
2427    Description:
2428    
2429    NEW BOOTFILES!!!    Version number bumped to 110.39.3.
2430    
2431    Primops have changed. This means that the bin/boot-file formats have
2432    changed as well.
2433    
2434    To make sure that there is no confusion, I made a new version.
2435    
2436    
2437    CHANGES:
2438    
2439    * removed REMT from mltree (remainder should never overflow).
2440    
2441    * added primops to deal with divisions of all flavors to the frontend
2442    
2443    * handled these primops all the way through so they map to their respective
2444      MLRISC support
2445    
2446    * used these primops in the implementation of Int, Int32, Word, Word32
2447    
2448    * removed INLDIV, INLMOD, and INLREM as they are no longer necessary
2449    
2450    * parameterized INLMIN, INLMAX, and INLABS by a numkind
2451    
2452    * translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
2453      floating point
2454    
2455    * used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
2456      and Real (but Real.abs maps to a separate floating-point-only primop)
2457    
2458    
2459    TODO items:
2460    
2461    * Hacked Alpha32 instruction selection, disabling the selection of REMx
2462      instructions because the machine instruction encoder cannot handle
2463      them.  (Hppa, PPC, and Sparc instruction selection did not handle
2464      REM in the first place, and REM is supported by the x86 machine coder.)
2465    
2466    * Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
2467      selection phase.  (The two can be streamlined because the hardware
2468      delivers both quotient and remainder at the same time anyway.)
2469    
2470    * Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
2471      (Currently the behavior is inconsistent both across architectures and
2472      wrt. the draft Basis spec.)
2473    
2474    * Word8 should eventually be handled natively, too.
2475    
2476    * There seems to be one serious bug in mltree-gen.sml.  It appears, though,
2477      as if there currently is no execution path that could trigger it in
2478      SML/NJ.  (The assumptions underlying functions arith and promotable do not
2479      hold for things like multiplication and division.)
2480    
2481    ----------------------------------------------------------------------
2482    Name: Matthias Blume
2483    Date: 2002/03/27 16:27:00 EST
2484    Tag: blume-20020327-mlrisc-divisions
2485    Description:
2486    
2487    Added support for all four division operations (ML's div, mod, quot,
2488    and rem) to MLRISC.  In the course of doing so, I also rationalized
2489    the naming (no more annoying switch-around of DIV and QUOT), by
2490    parameterizing the operation by div_rounding_mode (which can be either
2491    DIV_TO_ZERO or DIV_TO_NEGINF).
2492    
2493    The generic MLTreeGen functor takes care of compiling all four
2494    operations down to only round-to-zero div.
2495    
2496    Missing pieces:
2497    
2498      * Doing something smarter than relying on MLTreeGen on architectures
2499        like, e.g., the x86 where hardware division delivers both quotient and
2500        remainder at the same time.  With this, the implementation of the
2501        round-to-neginf operations could be further streamlined.
2502    
2503      * Remove inlining support for div/mod/rem from the frontend and replace it
2504        with primops that get carried through to the backend.  Do this for all
2505        int and word types.
2506    
2507    ----------------------------------------------------------------------
2508    Name: Matthias Blume
2509    Date: 2002/03/25 17:25:00 EST
2510    Tag: blume-20020325-divmod
2511    Description:
2512    
2513    I improved (hopefully without breaking them) the implementation of Int.div,
2514    Int.mod, and Int.rem.   For this, the code in translate.sml now takes
2515    advantage of the following observations:
2516    
2517      Let  q = x quot y      r = x rem y
2518           d = x div  y      m = x mod y
2519    
2520    where "quot" is the round-to-zero version of integer division that
2521    hardware usually provides.  Then we have:
2522    
2523         r = x - q * y        where neither the * nor the - will overflow
2524         d = if q >= 0 orelse x = q * y then q else q - 1
2525                              where neither the * nor the - will overflow
2526         m = if q >= 0 orelse r = 0 then r else r + y
2527                              where the + will not overflow
2528    
2529    This results in substantial simplification of the generated code.
2530    The following table shows the number of CFG nodes and edges generated
2531    for
2532            fun f (x, y) = x OPER y
2533            (* with OPER \in div, mod, quot, rem *)
2534    
2535    
2536        OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
2537        --------------------------------------------------------
2538         div |         24 |         39 |         12 |         16
2539         mod |         41 |         71 |         12 |         16
2540        quot |          8 |         10 |          8 |         10
2541         rem |         10 |         14 |          8 |         10
2542    
2543    
2544    ----------------------------------------------------------------------
2545    Name: Matthias Blume
2546    Date: 2002/03/25 22:06:00 EST
2547    Tag: blume-20020325-cprotobug
2548    Description:
2549    
2550    Fixed a bug in cproto (c prototype decoder).
2551    
2552    ----------------------------------------------------------------------
2553    Name: Matthias Blume
2554    Date: 2002/03/25 16:00:00 EST
2555    Tag: blume-20020325-raw-primops
2556    Description:
2557    
2558    I did some cleanup to Allen's new primop code and
2559    replaced yesterday's bootfiles with new ones.
2560    (But they are stored in the same place.)
2561    
2562    ----------------------------------------------------------------------
2563    Name: Matthias Blume
2564    Date: 2002/03/24 22:40:00 EST
2565    Tag: blume-20020324-bootfiles
2566    Description:
2567    
2568    Made the bootfiles that Allen asked for.
2569    
2570    ----------------------------------------------------------------------
2571    Name: Allen Leung
2572    Date: 2002/03/23 15:50:00 EST
2573    Tag: leunga-20020323-flint-cps-rcc-primops
2574    Description:
2575    
2576      1. Changes to FLINT primops:
2577    
2578        (* make a call to a C-function;
2579         * The primop carries C function prototype information and specifies
2580         * which of its (ML-) arguments are floating point. C prototype
2581         * information is for use by the backend, ML information is for
2582         * use by the CPS converter. *)
2583      | RAW_CCALL of { c_proto: CTypes.c_proto,
2584                       ml_args: ccall_type list,
2585                       ml_res_opt: ccall_type option,
2586                       reentrant : bool
2587                     } option
2588       (* Allocate uninitialized storage on the heap.
2589        * The record is meant to hold short-lived C objects, i.e., they
2590        * are not ML pointers.  With the tag, the representation is
2591        * the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
2592        *)
2593      | RAW_RECORD of {tag:bool,sz:int}
2594      and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
2595    
2596      2.  These CPS primops are now overloaded:
2597    
2598           rawload of {kind:numkind}
2599           rawstore of {kind:numkind}
2600    
2601          The one argument form is:
2602    
2603             rawload {kind} address
2604    
2605          The two argument form is:
2606    
2607             rawload {kind} [ml object, byte-offset]
2608    
2609      3. RAW_CCALL/RCC now takes two extra arguments:
2610    
2611         a. The first is whether the C call is reentrant, i.e., whether
2612            ML state should be saved and restored.
2613         b. The second argument is a string argument specifying the name of
2614            library and the C function.
2615    
2616         These things are currently not handled in the code generator, yet.
2617    
2618      4. In CProto,
2619    
2620         An encoding type of "bool" means "ml object" and is mapped into
2621         C prototype of PTR.  Note that "bool" is different than "string",
2622         even though "string" is also mapped into PTR, because "bool"
2623         is assigned an CPS type of BOGt, while "string" is assigned INT32t.
2624    
2625      5. Pickler/unpicker
2626    
2627         Changed to handle RAW_RECORD and newest RAW_CCALL
2628    
2629      6. MLRiscGen,
2630    
2631         1. Changed to handle the new rawload/rawstore/rawrecord operators.
2632         2. Code for handling C Calls has been moved to a new module CPSCCalls,
2633            in the file CodeGen/cpscompile/cps-c-calls.sml
2634    
2635      7. Added the conditional move operator
2636    
2637             condmove of branch
2638    
2639         to cps.  Generation of this is still buggy so it is currently
2640         disabled.
2641    
2642    ----------------------------------------------------------------------
2643    Name: Lal George
2644    Date: 2002/03/22 14:18:25 EST
2645    Tag: george-20020322-cps-branch-prob
2646    Description:
2647    
2648    Implemented the Ball-Larus branch prediction-heuristics, and
2649    incorporated graphical viewers for control flow graphs.
2650    
2651    Ball-Larus Heuristics:
2652    ---------------------
2653    See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
2654    
2655    By design it uses the Dempster-Shafer theory for combining
2656    probabilities.  For example, in the function:
2657    
2658        fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
2659    
2660    the ball-larus heuristics predicts that the n=0 is unlikely
2661    (OH-heuristic), and the 'then' branch is unlikely because of the
2662    RH-heuristic -- giving the 'then' branch an even lower combined
2663    probability using the Dempster-Shafer theory.
2664    
2665    Finally, John Reppy's loop analysis in MLRISC, further lowers the
2666    probability of the 'then' branch because of the loop in the else
2667    branch.
2668    
2669    
2670    Graphical Viewing:
2671    ------------------
2672    I merely plugged in Allen's graphical viewers into the compiler. The
2673    additional code is not much. At the top level, saying:
2674    
2675            Control.MLRISC.getFlag "cfg-graphical-view" := true;
2676    
2677    will display the graphical view of the control flow graph just before
2678    back-patching.  daVinci must be in your path for this to work. If
2679    daVinci is not available, then the default viewer can be changed
2680    using:
2681    
2682            Control.MLRISC.getString "viewer"
2683    
2684    which can be set to "dot" or "vcg" for the corresponding viewers. Of
2685    course, these viewers must be in your path.
2686    
2687    The above will display the compilation unit at the level of clusters,
2688    many of which are small, boring, and un-interesting. Also setting:
2689    
2690            Control.MLRISC.getInt "cfg-graphical-view_size"
2691    
2692    will display clusters that are larger than the value set by the above.
2693    
2694    
2695    ----------------------------------------------------------------------
2696    Name: Matthias Blume
2697    Date: 2002/03/21 22:20:00 EST
2698    Tag: blume-20020321-kmp-bugfix
2699    Description:
2700    
2701    Changed the interface to the KMP routine in PreString and fixed
2702    a minor bug in one place where it was used.
2703    
2704    ----------------------------------------------------------------------
2705    Name: Allen Leung
2706    Date: 2002/03/21 20:30:00 EST
2707    Tag: leunga-20020321-cfg
2708    Description:
2709    
2710      Fixed a potential problem in cfg edge splitting.
2711    
2712    ----------------------------------------------------------------------
2713    Name: Allen Leung
2714    Date: 2002/03/21 17:15:00 EST
2715    Tag: leunga-20020321-x86-fp-cfg
2716    Description:
2717    
2718      1. Recoded the buggy parts of x86-fp.
2719    
2720         a. All the block reordering code has been removed.
2721            We now depend on the block placement phases to do this work.
2722    
2723         b. Critical edge splitting code has been simplified and moved into the
2724            CFG modules, as where they belong.
2725    
2726         Both of these were quite buggy and complex.  The code is now much, much
2727         simpler.
2728    
2729      2. X86 backend.
2730    
2731         a. Added instructions for 64-bit support.  Instruction selection for
2732            64-bit has not been committed, however, since that
2733            requires changes to MLTREE which haven't been approved by
2734            Lal and John.
2735    
2736         b. Added support for FUCOMI and FUCOMIP when generating code for
2737            PentiumPro and above.  We only generate these instructions in
2738            the fast-fp mode.
2739    
2740         c. Added cases for JP and JNP in X86FreqProps.
2741    
2742      3. CFG
2743    
2744         CFG now has a bunch of methods for edge splitting and merging.
2745    
2746      4. Machine description.
2747    
2748         John's simplification of MLTREE_BASIS.fcond broke a few machine
2749         description things:
2750    
2751         rtl-build.{sig,sml} and hppa.mdl fixed.
2752    
2753         NOTE: the machine description stuff in the repository is still broken.
2754               Again, I can't put my fixes in because that involves
2755               changes to MLTREE.
2756    
2757    ----------------------------------------------------------------------
2758    Name: Matthias Blume
2759    Date: 2002/03/20 15:55:00 EST
2760    Tag: blume-20020320-kmp
2761    Description:
2762    
2763    Implemented Knuth-Morris-Pratt string matching in PreString and used
2764    it for String.isSubstring, Substring.isSubstring, and
2765    Substring.position.
2766    
2767    (Might need some stress-testing.  Simple examples worked fine.)
2768    
2769    ----------------------------------------------------------------------
2770    Name: Matthias Blume
2771    Date: 2002/03/19 16:37:00 EST
2772    Tag: blume-20020319-witnesses
2773    Description:
2774    
2775    Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
2776    
2777    This implements a generic mechanism for changing constness qualifiers
2778    anywhere within big C types without resorting to outright "casts".
2779    (So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
2780    modify the constness at the outermost level.)
2781    The implementation of "convert" is based on the idea of "witness"
2782    values -- values that are not used by the operation but whose types
2783    "testify" to their applicability.  On the implementation side, "convert"
2784    is simply a projection (returning its second curried argument).  With
2785    cross-module inlining, it should not result in any machine code being
2786    generated.
2787    
2788    ----------------------------------------------------------------------
2789    Name: Matthias Blume
2790    Date: 2002/03/15 16:40:00 EST
2791    Tag: blume-20020315-basis
2792    Description:
2793    
2794    Provided (preliminary?) implementations for
2795    
2796      {String,Substring}.{concatWith,isSuffix,isSubstring}
2797    
2798    and
2799    
2800      Substring.full
2801    
2802    Those are in the Basis spec but they were missing in SML/NJ.
2803    
2804    ----------------------------------------------------------------------
2805    Name: Matthias Blume
2806    Date: 2002/03/14 21:30:00 EST
2807    Tag: blume-20020314-controls
2808    Description:
2809    
2810    Controls:
2811    ---------
2812    
2813    1. Factored out the recently-added Controls : CONTROLS stuff and put
2814       it into its own library $/controls-lib.cm.  The source tree for
2815       this is under src/smlnj-lib/Controls.
2816    
2817    2. Changed the names of types and functions in this interface, so they
2818       make a bit more "sense":
2819    
2820          module -> registry
2821          'a registry -> 'a group
2822    
2823    3. The interface now deals in ref cells only.  The getter/setter interface
2824       is (mostly) gone.
2825    
2826    4. Added a function that lets one register an already-existing ref cell.
2827    
2828    5. Made the corresponding modifications to the rest of the code so that
2829       everything compiles again.
2830    
2831    6. Changed the implementation of Controls.MLRISC back to something closer
2832       to the original.  In particular, this module (and therefore MLRISC)
2833       does not depend on Controls.  There now is some link-time code in
2834       int-sys.sml that registers the MLRISC controls with the Controls
2835       module.
2836    
2837    CM:
2838    ---
2839    
2840      * One can now specify the lambda-split aggressiveness in init.cmi.
2841    
2842    ----------------------------------------------------------------------
2843    Name: Allen Leung
2844    Date: 2002/03/13 17:30:00 EST
2845    Tag: leunga-20020313-x86-fp-unary
2846    Description:
2847    
2848    Bug fix for:
2849    
2850    > leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
2851    > Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
2852    > - fun f(x,(y,z)) = Real.~ y;
2853    > [autoloading]
2854    > [autoloading done]
2855    >       fchsl   (%eax), 184(%esp)
2856    > Error: MLRisc bug: X86MCEmitter.emitInstr
2857    >
2858    > uncaught exception Error
2859    >   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
2860    
2861    The problem was that the code generator did not generate any fp registers
2862    in this case, and the ra didn't know that it needed to run the X86FP phase to
2863    translate the pseudo fp instruction.   This only happened with unary fp
2864    operators in certain situations.
2865    
2866    ----------------------------------------------------------------------
2867    Name: Matthias Blume
2868    Date: 2002/03/13 14:00:00 EST
2869    Tag: blume-20020313-overload-etc
2870    Description:
2871    
2872    1. Added _overload as a synonym for overload for backward compatibility.
2873       (Control.overloadKW must be true for either version to be accepted.)
2874    
2875    2. Fixed bug in install script that caused more things to be installed
2876       than what was requested in config/targets.
2877    
2878    3. Made CM aware of the (_)overload construct so that autoloading
2879       works.
2880    
2881    ----------------------------------------------------------------------
2882    Name: Matthias Blume
2883    Date: 2002/03/12 22:03:00 EST
2884    Tag: blume-20020312-url
2885    Description:
2886    
2887    Forgot to update BOOT and srcarchiveurl.
2888    
2889    ----------------------------------------------------------------------
2890    Name: Matthias Blume
2891    Date: 2002/03/12 17:30:00 EST
2892    Tag: blume-20020312-version110392
2893    Description:
2894    
2895    Yet another version number bump (because of small changes to the
2896    binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
2897    
2898    Changes:
2899    
2900      The new pid generation scheme described a few weeks ago was overly
2901      complicated.  I implemented a new mechanism that is simpler and
2902      provides a bit more "stability":  Once CM has seen a compilation
2903      unit, it keeps its identity constant (as long as you do not delete
2904      those crucial CM/GUID/* files).  This means that when you change
2905      an interface, compile, then go back to the old interface, and
2906      compile again, you arrive at the original pid.
2907    
2908      There now also is a mechanism that instructs CM to use the plain
2909      environment hash as a module's pid (effectively making its GUID
2910      the empty string).  For this, "noguid" must be specified as an
2911      option to the .sml file in question within its .cm file.
2912      This is most useful for code that is being generated by tools such
2913      as ml-nlffigen (because during development programmers tend to
2914      erase the tool's entire output directory tree including CM's cached
2915      GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
2916      revert to the old, broken behavior of SML/NJ, but in specific cases
2917      where there is no danger of interface confusion, its use is ok
2918      (I think).
2919    
2920      ml-nlffigen by default generates "noguid" annotations.  They can be
2921      turned off by specifying -guid in its command line.
2922    
2923    ----------------------------------------------------------------------
2924    Name: Lal George
2925    Date: 2002/03/12 12 14:42:36 EST
2926    Tag: george-20020312-frequency-computation
2927    Description:
2928    
2929    Integrated jump chaining and static block frequency into the
2930    compiler. More details and numbers later.
2931    
2932    ----------------------------------------------------------------------
2933    Name: Lal George
2934    Date: 2002/03/11 11 22:38:53 EST
2935    Tag: george-20020311-jump-chain-elim
2936    Description:
2937    
2938    Tested the jump chain elimination on all architectures (except the
2939    hppa).  This is on by default right now and is profitable for the
2940    alpha and x86, however, it may not be profitable for the sparc and ppc
2941    when compiling the compiler.
2942    
2943    The gc test will typically jump to a label at the end of the cluster,
2944    where there is another jump to an external cluster containing the actual
2945    code to invoke gc. This is to allow factoring of common gc invocation
2946    sequences. That is to say, we generate:
2947    
2948            f:
2949               testgc
2950               ja   L1      % jump if above to L1
2951    
2952            L1:
2953               jmp L2
2954    
2955    
2956    After jump chain elimination the 'ja L1' instructions is converted to
2957    'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
2958    up being implemented in their long form (if L2 is far away) using:
2959    
2960            jbe     L3      % jump if below or equal to L3
2961            jmp     L2
2962         L3:
2963            ...
2964    
2965    
2966    For large compilation units L2  may be far away.
2967    
2968    
2969    ----------------------------------------------------------------------
2970    Name: Matthias Blume
2971    Date: 2002/03/11 13:30:00 EST
2972    Tag: blume-20020311-mltreeeval
2973    Description:
2974    
2975    A functor parameter was missing.
2976    
2977    ----------------------------------------------------------------------
2978    Name: Allen Leung
2979    Date: 2002/03/11 10:30:00 EST
2980    Tag: leunga-20020311-runtime-string0
2981    Description:
2982    
2983       The representation of the empty string now points to a
2984    legal null terminated C string instead of unit.  It is now possible
2985    to convert an ML string into C string with InlineT.CharVector.getData.
2986    This compiles into one single machine instruction.
2987    
2988    ----------------------------------------------------------------------
2989    Name: Allen Leung
2990    Date: 2002/03/10 23:55:00 EST
2991    Tag: leunga-20020310-x86-call
2992    Description:
2993    
2994       Added machine generation for CALL instruction (relative displacement mode)
2995    
2996    ----------------------------------------------------------------------
2997    Name: Matthias Blume
2998    Date: 2002/03/08 16:05:00
2999    Tag: blume-20020308-entrypoints
3000    Description:
3001    
3002    Version number bumped to 110.39.1.  NEW BOOTFILES!
3003    
3004    Entrypoints: non-zero offset into a code object where execution should begin.
3005    
3006    - Added the notion of an entrypoint to CodeObj.
3007    - Added reading/writing of entrypoint info to Binfile.
3008    - Made runtime system bootloader aware of entrypoints.
3009    - Use the address of the label of the first function given to mlriscGen
3010      as the entrypoint.  This address is currently always 0, but it will
3011      not be 0 once we turn on block placement.
3012    - Removed the linkage cluster code (which was The Other Way(tm) of dealing
3013      with entry points) from mlriscGen.
3014    
3015    ----------------------------------------------------------------------
3016    Name: Allen Leung
3017    Date: 2002/03/07 20:45:00 EST
3018    Tag: leunga-20020307-x86-cmov
3019    Description:
3020    
3021       Bug fixes for CMOVcc on x86.
3022    
3023       1. Added machine code generation for CMOVcc
3024       2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
3025       3. CMOVcc cannot have an immediate operand as argument.
3026    
3027    ----------------------------------------------------------------------
3028    Name: Matthias Blume
3029    Date: 2002/03/07 16:15:00 EST
3030    Tag: blume-20020307-controls
3031    Description:
3032    
3033    This is a very large but mostly boring patch which makes (almost)
3034    every tuneable compiler knob (i.e., pretty much everything under
3035    Control.* plus a few other things) configurable via both the command
3036    line and environment variables in the style CM did its configuration
3037    until now.
3038    
3039    Try starting sml with '-h' (or, if you are brave, '-H')
3040    
3041    To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
3042    implements the underlying generic mechanism.
3043    
3044    The interface to some of the existing such facilities has changed somewhat.
3045    For example, the MLRiscControl module now provides mkFoo instead of getFoo.
3046    (The getFoo interface is still there for backward-compatibility, but its
3047    use is deprecated.)
3048    
3049    The ml-build script passes -Cxxx=yyy command-line arguments through so
3050    that one can now twiddle the compiler settings when using this "batch"
3051    compiler.
3052    
3053    TODO items:
3054    
3055    We should go through and throw out all controls that are no longer
3056    connected to anything.  Moreover, we should go through and provide
3057    meaningful (and correct!) documentation strings for those controls
3058    that still are connected.
3059    
3060    Currently, multiple calls to Controls.new are accepted (only the first
3061    has any effect).  Eventually we should make sure that every control
3062    is being made (via Controls.new) exactly once.  Future access can then
3063    be done using Controls.acc.
3064    
3065    Finally, it would probably be a good idea to use the getter-setter
3066    interface to controls rather than ref cells.  For the time being, both
3067    styles are provided by the Controls module, but getter-setter pairs are
3068    better if thread-safety is of any concern because they can be wrapped.
3069    
3070    *****************************************
3071    
3072    One bug fix: The function blockPlacement in three of the MLRISC
3073    backpatch files used to be hard-wired to one of two possibilities at
3074    link time (according to the value of the placementFlag).  But (I
3075    think) it should rather sense the flag every time.
3076    
3077    *****************************************
3078    
3079    Other assorted changes (by other people who did not supply a HISTORY entry):
3080    
3081    1. the cross-module inliner now works much better (Monnier)
3082    2. representation of weights, frequencies, and probabilities in MLRISC
3083       changed in preparation of using those for weighted block placement
3084       (Reppy, George)
3085    
3086    ----------------------------------------------------------------------
3087    Name: Lal George
3088    Date: 2002/03/07 14:44:24 EST 2002
3089    Tag: george-20020307-weighted-block-placement
3090    
3091    Tested the weighted block placement optimization on all architectures
3092    (except the hppa) using AMPL to generate the block and edge frequencies.
3093    Changes were required in the machine properties to correctly
3094    categorize trap instructions. There is an MLRISC flag
3095    "weighted-block-placement" that can be used to enable weighted block
3096    placement, but this will be ineffective without block/edge
3097    frequencies (coming soon).
3098    
3099    
3100    ----------------------------------------------------------------------
3101    Name: Lal George
3102    Date: 2002/03/05 17:24:48 EST
3103    Tag: george-20020305-linkage-cluster
3104    
3105    In order to support the block placement optimization, a new cluster
3106    is generated as the very first cluster (called the linkage cluster).
3107    It contains a single jump to the 'real' entry point for the compilation
3108    unit. Block placement has no effect on the linkage cluster itself, but
3109    all the other clusters  have full freedom in the manner in which they
3110    reorder blocks or functions.
3111    
3112    On the x86 the typical linkage code that is generated is:
3113       ----------------------
3114            .align 2
3115       L0:
3116            addl    $L1-L0, 72(%esp)
3117            jmp     L1
3118    
3119    
3120            .align  2
3121       L1:
3122       ----------------------
3123    
3124    72(%esp) is the memory location for the stdlink register. This
3125    must contain the address of the CPS function being called. In the
3126    above example, it contains the address of  L0; before
3127    calling L1 (the real entry point for the compilation unit), it
3128    must contain the address for L1, and hence
3129    
3130            addl $L1-L0, 72(%esp)
3131    
3132    I have tested this on all architectures except the hppa.The increase
3133    in code size is of course negligible
3134    
3135    ----------------------------------------------------------------------
3136    Name: Allen Leung
3137    Date: 2002/03/03 13:20:00 EST
3138    Tag: leunga-20020303-mlrisc-tools
3139    
3140      Added #[ ... ] expressions to mlrisc tools
3141    
3142    ----------------------------------------------------------------------
3143    Name: Matthias Blume
3144    Date: 2002/02/27 12:29:00 EST
3145    Tag: blume-20020227-cdebug
3146    Description:
3147    
3148    - made types in structure C and C_Debug to be equal
3149    - got rid of code duplication (c-int.sml vs. c-int-debug.sml)
3150    - there no longer is a C_Int_Debug (C_Debug is directly derived from C)
3151    
3152    ----------------------------------------------------------------------
3153    Name: Matthias Blume
3154    Date: 2002/02/26 12:00:00 EST
3155    Tag: blume-20020226-ffi
3156    Description:
3157    
3158    1. Fixed a minor bug in CM's "noweb" tool:
3159       If numbering is turned off, then truly don't number (i.e., do not
3160       supply the -L option to noweb).  The previous behavior was to supply
3161       -L'' -- which caused noweb to use the "default" line numbering scheme.
3162       Thanks to Chris Richards for pointing this out (and supplying the fix).
3163    
3164    2. Once again, I reworked some aspects of the FFI:
3165    
3166       A. The incomplete/complete type business:
3167    
3168       - Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
3169         gone!
3170       - ML types representing an incomplete type are now *equal* to
3171         ML types representing their corresponding complete types (just like
3172         in C).  This is still safe because ml-nlffigen will not generate
3173         RTTI for incomplete types, nor will it generate functions that
3174         require access to such RTTI.   But when ML code generated from both
3175         incomplete and complete versions of the C type meet, the ML types
3176         are trivially interoperable.
3177    
3178         NOTE:  These changes restore the full generality of the translation
3179         (which was previously lost when I eliminated functorization)!
3180    
3181       B. Enum types:
3182    
3183       - Structure C now has a type constructor "enum" that is similar to
3184         how the "su" constructor works.  However, "enum" is not a phantom
3185         type because each "T enum" has values (and is isomorphic to
3186         MLRep.Signed.int).
3187       - There are generic access operations for enum objects (using
3188         MLRep.Signed.int).
3189       - ml-nlffigen will generate a structure E_foo for each "enum foo".
3190         * The structure contains the definition of type "mlrep" (the ML-side
3191         representation type of the enum).  Normally, mlrep is the same
3192         as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
3193         then mlrep will be defined as a datatype -- thus facilitating
3194         pattern matching on mlrep values.
3195         ("-ec" will be suppressed if there are duplicate values in an
3196          enumeration.)
3197         * Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
3198         will be generated for each C enum constant xxx.
3199         * Conversion functions m2i and i2m convert between mlrep and
3200         MLRep.Signed.int.  (Without "-ec", these functions are identities.)
3201         * Coversion functions c and ml convert between mlrep and "tag enum".
3202         * Access functions (get/set) fetch and store mlrep values.
3203       - By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
3204         enumerations are merged into one single enumeration represented by
3205         structure E_'.
3206    
3207    ----------------------------------------------------------------------
3208    Name: Allen Leung
3209    Date: 2002/02/25 04:45:00 EST
3210    Tag: leunga-20020225-cps-spill
3211    
3212    This is a new implementation of the CPS spill phase.
3213    The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
3214    In case of problems, replace it with the old file spill.sml
3215    
3216    The current compiler runs into some serious performance problems when
3217    constructing a large record.  This can happen when we try to compile a
3218    structure with many items.  Even a very simple structure like the following
3219    makes the compiler slow down.
3220    
3221        structure Foo = struct
3222           val x_1 = 0w1 : Word32.int
3223           val x_2 = 0w2 : Word32.int
3224           val x_3 = 0w3 : Word32.int
3225           ...
3226           val x_N = 0wN : Word32.int
3227        end
3228    
3229    The following table shows the compile time, from N=1000 to N=4000,
3230    with the old compiler:
3231    
3232    N
3233    1000   CPS 100 spill                           0.04u  0.00s  0.00g
3234           MLRISC ra                               0.06u  0.00s  0.05g
3235              (spills = 0 reloads = 0)
3236           TOTAL                                   0.63u  0.07s  0.21g
3237    
3238    1100   CPS 100 spill                           8.25u  0.32s  0.64g
3239           MLRISC ra                               5.68u  0.59s  3.93g
3240              (spills = 0 reloads = 0)
3241           TOTAL                                   14.71u  0.99s  4.81g
3242    
3243    1500   CPS 100 spill                           58.55u  2.34s  1.74g
3244           MLRISC ra                               5.54u  0.65s  3.91g
3245              (spills = 543 reloads = 1082)
3246           TOTAL                                   65.40u  3.13s  6.00g
3247    
3248    2000   CPS 100 spill                           126.69u  4.84s  3.08g
3249           MLRISC ra                               0.80u  0.10s  0.55g
3250              (spills = 42 reloads = 84)
3251           TOTAL                                   129.42u  5.10s  4.13g
3252    
3253    3000   CPS 100 spill                           675.59u  19.03s  11.64g
3254           MLRISC ra                               2.69u  0.27s  1.38g
3255              (spills = 62 reloads = 124)
3256           TOTAL                                   682.48u  19.61s  13.99g
3257    
3258    4000   CPS 100 spill                           2362.82u  56.28s  43.60g
3259           MLRISC ra                               4.96u  0.27s  2.72g
3260              (spills = 85 reloads = 170)
3261           TOTAL                                   2375.26u  57.21s  48.00g
3262    
3263    As you can see the old cps spill module suffers from some serious
3264    performance problem.  But since I cannot decipher the old code fully,
3265    instead of patching the problems up, I'm reimplementing it
3266    with a different algorithm.  The new code is more modular,
3267    smaller when compiled, and substantially faster
3268    (O(n log n) time and O(n) space).  Timing of the new spill module:
3269    
3270    4000  CPS 100 spill                           0.02u  0.00s  0.00g
3271          MLRISC ra                               0.25u  0.02s  0.15g
3272             (spills=1 reloads=3)
3273          TOTAL                                   7.74u  0.34s  1.62g
3274    
3275    Implementation details:
3276    
3277    As far as I can tell, the purpose of the CPS spill module is to make sure the
3278    number of live variables at any program point (the bandwidth)
3279    does not exceed a certain limit, which is determined by the
3280    size of the spill area.
3281    
3282    When the bandwidth is too large, we decrease the register pressure by
3283    packing live variables into spill records.  How we achieve this is
3284    completely different than what we did in the old code.
3285    
3286    First, there is something about the MLRiscGen code generator
3287    that we should be aware of:
3288    
3289    o MLRiscGen performs code motion!
3290    
3291       In particular, it will move floating point computations and
3292       address computations involving only the heap pointer to
3293       their use sites (if there is only a single use).
3294       What this means is that if we have a CPS record construction
3295       statement
3296    
3297           RECORD(k,vl,w,e)
3298    
3299       we should never count the new record address w as live if w
3300       has only one use (which is often the case).
3301    
3302       We should do something similar to floating point, but the transformation
3303       there is much more complex, so I won't deal with that.
3304    
3305    Secondly, there are now two new cps primops at our disposal:
3306    
3307     1. rawrecord of record_kind option
3308        This pure operator allocates some uninitialized storage from the heap.
3309        There are two forms:
3310    
3311         rawrecord NONE [INT n]  allocates a tagless record of length n
3312         rawrecord (SOME rk) [INT n] allocates a tagged record of length n
3313                                     and initializes the tag.
3314    
3315     2. rawupdate of cty
3316          rawupdate cty (v,i,x)
3317          Assigns to x to the ith component of record v.
3318          The storelist is not updated.
3319    
3320    We use these new primops for both spilling and increment record construction.
3321    
3322     1. Spilling.
3323    
3324        This is implemented with a linear scan algorithm (but generalized
3325        to trees).  The algorithm will create a single spill record at the
3326        beginning of the cps function and use rawupdate to spill to it,
3327        and SELECT or SELp to reload from it.  So both spills and reloads
3328        are fine-grain operations.  In contrast, in the old algorithm
3329        "spills" have to be bundled together in records.
3330    
3331        Ideally, we should sink the spill record construction to where
3332        it is needed.  We can even split the spill record into multiple ones
3333        at the places where they are needed.  But CPS is not a good
3334        representation for global code motion, so I'll keep it simple and
3335        am not attempting this.
3336    
3337     2. Incremental record construction (aka record splitting).
3338    
3339        Long records with many component values which are simulatenously live
3340        (recall that single use record addresses are not considered to
3341         be live) are constructed with rawrecord and rawupdate.
3342        We allocate space on the heap with rawrecord first, then gradually
3343        fill it in with rawupdate.  This is the technique suggested to me
3344        by Matthias.
3345    
3346        Some restrictions on when this is applicable:
3347        1. It is not a VECTOR record.  The code generator currently does not handle
3348           this case. VECTOR record uses double indirection like arrays.
3349        2. All the record component values are defined in the same "basic block"
3350           as the record constructor.  This is to prevent speculative
3351           record construction.
3352    
3353    ----------------------------------------------------------------------
3354    Name: Allen Leung
3355    Date: 2002/02/22 01:02:00 EST
3356    Tag: leunga-20020222-mlrisc-tools
3357    
3358    Minor bug fixes in the parser and rewriter
3359    
3360    ----------------------------------------------------------------------
3361    Name: Allen Leung
3362    Date: 2002/02/21 20:20:00 EST
3363    Tag: leunga-20020221-peephole
3364    
3365    Regenerated the peephole files.  Some contained typos in the specification
3366    and some didn't compile because of pretty printing bugs in the old version
3367    of 'nowhere'.
3368    
3369    ----------------------------------------------------------------------
3370    Name: Allen Leung
3371    Date: 2002/02/19 20:20:00 EST
3372    Tag: leunga-20020219-mlrisc-tools
3373    Description:
3374    
3375       Minor bug fixes to the mlrisc-tools library:
3376    
3377       1.  Fixed up parsing colon suffixed keywords
3378       2.  Added the ability to shut the error messages up
3379       3.  Reimplemented the pretty printer and fixed up/improved
3380           the pretty printing of handle and -> types.
3381       4.  Fixed up generation of literal symbols in the nowhere tool.
3382       5.  Added some SML keywords to to sml.sty
3383    
3384    ----------------------------------------------------------------------
3385    Name: Matthias Blume
3386    Date: 2002/02/19 16:20:00 EST
3387    Tag: blume-20020219-cmffi
3388    Description:
3389    
3390    A wild mix of changes, some minor, some major:
3391    
3392    * All C FFI-related libraries are now anchored under $c:
3393        $/c.cm      --> $c/c.cm
3394        $/c-int.cm  --> $c/internals/c-int.cm
3395        $/memory.cm --> $c/memory/memory.cm
3396    
3397    * "make" tool (in CM) now treats its argument pathname slightly
3398      differently:
3399        1. If the native expansion is an absolute name, then before invoking
3400           the "make" command on it, CM will apply OS.Path.mkRelative
3401           (with relativeTo = OS.FileSys.getDir()) to it.
3402        2. The argument will be passed through to subsequent phases of CM
3403           processing without "going native".  In particular, if the argument
3404           was an anchored path, then "make" will not lose track of that anchor.
3405    
3406    * Compiler backends now "know" their respective C calling conventions
3407      instead of having to be told about it by ml-nlffigen.  This relieves
3408      ml-nlffigen from one of its burdens.
3409    
3410    * The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
3411    
3412    * Export C_DEBUG and C_Debug from $c/c.cm.
3413    
3414    * C type encoding in ml-nlffi-lib has been improved to model the conceptual
3415      subtyping relationship between incomplete pointers and their complete
3416      counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
3417      with the convention of instantiating 'o with ('t, 'c) obj whenever
3418      the pointer target type is complete.  In the incomplete case, 'o
3419      will be instantiated with some "'c iobj" -- a type obtained by
3420      using one of the functors PointerToIncompleteType or PointerToCompleteType.
3421    
3422      Operations that work on both incomplete and complete pointer types are
3423      typed as taking an 'o ptr while operations that require the target to
3424      be known are typed as taking some ('t, 'c) obj ptr.
3425    
3426      voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
3427      where void is an eqtype without any values.  This makes it possible
3428      to work on voidptr values using functions meant to operate on light
3429      incomplete pointers.
3430    
3431    * As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
3432      been vastly simplified.
3433    
3434    ----------------------------------------------------------------------
3435    Name: Matthias Blume
3436    Date: 2002/02/19 10:48:00 EST
3437    Tag: blume-20020219-pqfix
3438    Description:
3439    
3440    Applied Chris Okasaki's bug fix for priority queues.
3441    
3442    ----------------------------------------------------------------------
3443    Name: Matthias Blume
3444    Date: 2002/02/15 17:05:00
3445    Tag: Release_110_39
3446    Description:
3447    
3448    Last-minute retagging is becoming a tradition... :-(
3449    
3450    This is the working release 110.39.
3451    
3452    ----------------------------------------------------------------------
3453    Name: Matthias Blume
3454    Date: 2002/02/15 16:00:00 EST
3455    Tag: Release_110_39-orig
3456    Description:
3457    
3458    Working release 110.39.  New bootfiles.
3459    
3460    (Update: There was a small bug in the installer so it wouldn't work
3461    with all shells.  So I retagged. -Matthias)
3462    
3463    ----------------------------------------------------------------------
3464    Name: Matthias Blume
3465    Date: 2002/02/15 14:17:00 EST
3466    Tag: blume-20020215-showbindings
3467    Description:
3468    
3469    Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
3470    the latter can be useful for exploring what bindings are available at
3471    the interactive prompt.  (The first function returns only the list
3472    of symbols that are really bound, the second prints those but also the
3473    ones that CM's autoloading mechanism knows about.)
3474    
3475    ----------------------------------------------------------------------
3476    Name: Matthias Blume
3477    Date: 2002/02/15 12:08:00 EST
3478    Tag: blume-20020215-iptrs
3479    Description:
3480    
3481    Two improvements to ml-nlffigen:
3482    
3483      1. Write files only if they do not exist or if their current contents
3484         do not coincide with what's being written.  (That is, avoid messing
3485         with the time stamps unless absolutely necessary.)
3486    
3487      2. Implement a "repository" mechanism for generated files related
3488         to "incomplete pointer types".   See the README file for details.
3489    
3490    ----------------------------------------------------------------------
3491    Name: Matthias Blume
3492    Date: 2002/02/14 11:50:00 EST
3493    Tag: blume-20020214-quote
3494    Description:
3495    
3496    Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
3497    because of the new and improved tag generation scheme.  (Thanks to Allen
3498    Leung for pointing it out.)
3499    
3500    ----------------------------------------------------------------------
3501    Name: Lal George
3502    Date: 2002/02/14 09:55:27 EST 2002
3503    Tag: george-20020214-isabelle-bug
3504    Description:
3505    
3506    Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
3507    of Isabelle on the x86.
3508    
3509    From Allen:
3510    -----------
3511     I've found the problem:
3512    
3513         in ra-core.sml, I use the counter "blocked" to keep track of the
3514         true number of elements in the freeze queue.  When the counter goes
3515         to zero, I skip examining the queue.  But I've messed up the
3516         bookkeeping in combine():
3517    
3518             else ();
3519             case !ucol of
3520               PSEUDO => (if !cntv > 0 then
3521                     (if !cntu > 0 then blocked := !blocked - 1 else ();
3522                                        ^^^^^^^^^^^^^^^^^^^^^^^
3523                      moveu := mergeMoveList(!movev, !moveu)
3524                     )
3525                  else ();
3526    
3527         combine() is called to coalesce two nodes u and v.
3528         I think I was thinking that if the move counts of u and v are both
3529         greater than zero then after they are coalesced then one node is
3530         removed from the freeze queue.  Apparently I was thinking that
3531         both u and v are of low degree, but that's clearly not necessarily true.
3532    
3533    
3534    02/12/2002:
3535        Here's the patch.  HOL now compiles.
3536    
3537        I don't know how this impact on performance (compile
3538        time or runtime).  This bug caused the RA (especially on the x86)
3539        to go thru the potential spill phase when there are still nodes on the
3540        freeze queue.
3541    
3542    
3543    
3544    
3545    ----------------------------------------------------------------------
3546    Name: Matthias Blume
3547    Date: 2002/02/13 22:40:00 EST
3548    Tag: blume-20020213-fptr-rtti
3549    Description:
3550    
3551    Fixed a bug in ml-nlffigen that was introduced with one of the previous
3552    updates.
3553    
3554    ----------------------------------------------------------------------
3555    Name: Matthias Blume
3556    Date: 2002/02/13 16:41:00 EST
3557    Tag: blume-20020213-cmlpq
3558    Description:
3559    
3560    Added new priority queue export symbols (which have just been added to
3561    smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
3562    not compile and the installer would choke.)
3563    
3564    ----------------------------------------------------------------------
3565    Name: Matthias Blume
3566    Date: 2002/02/13 16:15:00 EST
3567    Tag: blume-20020213-various
3568    Description:
3569    
3570    1. More tweaks to ml-nlffigen:
3571    
3572       - better internal datastructures (resulting in slight speedup)
3573       - "-match" option requires exact match
3574       - "localized" gensym counters (untagged structs/unions nested within
3575         other structs/unions or within typedefs get a fresh counter; their
3576         tag will be prefixed by a concatenation of their parents' tags)
3577       - bug fixes (related to calculation of transitive closure of types
3578         to be included in the output)
3579    
3580    2. Minor Basis updates:
3581    
3582       - added implementations for List.collate and Option.app
3583    
3584    ----------------------------------------------------------------------
3585    Name: Matthias Blume
3586    Date: 2002/02/11 15:55:00 EST
3587    Tag: blume-20020211-gensym
3588    Description:
3589    
3590    Added a "-gensym" option to command line of ml-nlffigen.  This can be
3591    used to specify a "stem" -- a string that is inserted in all "gensym'd"
3592    names (ML structure names that correspond to unnamed C structs, unions,
3593    and enums), so that separate runs of ml-nlffigen do not clash.
3594    
3595    ----------------------------------------------------------------------
3596    Name: Matthias Blume
3597    Date: 2002/02/11 12:05:00 EST
3598    Tag: blume-20020211-gensml
3599    Description:
3600    
3601    A quick fix for a problem with GenSML (in the pgraph-util library):
3602    Make generation of toplevel "local" optional.  (Strictly speaking,
3603    signature definitions within "local" are not legal SML.)
3604    
3605    Other than that: updates to INSTALL and cm/TODO.
3606    
3607    ----------------------------------------------------------------------
3608    Name: Matthias Blume
3609    Date: 2002/02/08 15:00:00 EST
3610    Tag: blume-20020208-uniquepid
3611    Description:
3612    
3613    0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
3614    
3615    1. The installer (config/install.sh) has gotten smarter:
3616    
3617         - Configuration options are a bit easier to specify now
3618           (in config/targets).
3619         - Bug in recognizing .tar.bz2 files fixed.
3620         - Installer automatically resolves dependencies between
3621           configuration options (e.g., if you ask for eXene, you will
3622           also get cml -- regardless whether you asked for it or not).
3623         - Installer can run in "quieter mode" by setting the environment
3624           variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
3625           "completely silent", though.
3626         - Build HashCons library as part of smlnj-lib.
3627    
3628    2. A new scheme for assigning persistent identifiers to compilation
3629       units (and, by extension, to types etc.) has been put into place.
3630       This fixes a long-standing bug where types and even dynamic values
3631       can get internally confused, thereby compromising type safety
3632       (abstraction) and dynamic correctness.  See
3633    
3634         http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
3635    
3636       for an example of how things could go wrong until now.
3637    
3638       The downside of the new scheme is that pids are not quite as
3639       persistent as they used to be: CM will generate a fresh pid
3640       for every compilation unit that it thinks it sees for the first
3641       time.  That means that if you compile starting from a clean, fresh
3642       source tree at two different times, you end up with different
3643       binaries.
3644    
3645       Cutoff recompilation, however, has not been compromised because
3646       CM keeps pid information in special caches between runs.
3647    
3648    ----------------------------------------------------------------------
3649    Name: Lal George
3650    Date: 2002/02/07 15:34:13 EST 2002
3651    Tag: <none>
3652    Description:
3653    
3654    Compilers that generate assembly code may produce  global labels
3655    whose value is resolved at link time. The various peephole optimization
3656    modules did not take this in account.
3657    
3658    TODO. The Labels.addrOf function should really return an option
3659    type so that clients are forced to deal with this issue, rather
3660    than an exception being raised.
3661    
3662    ----------------------------------------------------------------------
3663    Name: Lal George
3664    Date: 2002/02/06 13:55:02 EST
3665    Tag: george-20020206-ra-breakup
3666    Description:
3667    
3668    1. A bug fix from Allen.
3669    
3670        A typo causes extra fstp %st(0)'s to be generated at compensation
3671        edges, which might cause stack underflow traps at runtime.  This
3672        occurs in fft where there are extraneous fstps right before the 'into'
3673        trap instruction (in this case they are harmless since none of the
3674        integers overflow.)
3675    
3676    2. Pulled out various utility modules that were embedded in the modules
3677       of the register allocator. I need these modules for other purposes, but
3678       they are not complete enough to put into a library (just yet).
3679    ----------------------------------------------------------------------
3680    Name: Matthias Blume
3681    Date: 2002/01/31 16:05:00 EST
3682    Tag: blume-20020131-sparc-ccalls
3683    Description:
3684    
3685    1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
3686       of extra stack space by mistake.  Fixed.
3687    
3688    2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
3689    
3690    ----------------------------------------------------------------------
3691    Name: Allen Leung
3692    Date: 2002/01/30
3693    Tag: leunga-20020130-nowhere-bug-fix
3694    Description:
3695    
3696       MLRISC bug fixes:
3697       1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
3698       2. MachineInt.fromString was negating its value.
3699    
3700    ----------------------------------------------------------------------
3701    Name: Matthias Blume
3702    Date: 2002/01/29
3703    Tag: blume-20020129-INSTALL
3704    Description:
3705    
3706    - Added somewhat detailed installation instructions (file INSTALL).
3707    - Fixed curl-detection bug in config/install.sh.
3708    - It is now possible to select the URL getter using the URLGETTER
3709      environment variable:
3710    
3711          not set / "unknown"      --> automatic detection (script tries wget,
3712                                       curl, and lynx)
3713          "wget" / "curl" / "lynx" --> use the specified program (script "knows"
3714                                       how to properly invoke them)
3715          other                    --> use $URLGETTER directly, it must take
3716                                       precisely two command-line arguments
3717                                       (source URL and destination file name)
3718    
3719    ----------------------------------------------------------------------
3720    Name: Matthias Blume
3721    Date: 2002/01/28
3722    Tag: blume-20020128-sparc-ccalls
3723    Description:
3724    
3725    - Fixed problem with calculation of "used" registers in sparc-c-calls.
3726    - Make use of the allocParam argument in sparc-c-calls.
3727    
3728    ----------------------------------------------------------------------
3729    Name: Matthias Blume
3730    Date: 2002/01/28
3731    Tag: blume-20020128-allocParam
3732    Description:
3733    
3734    John Reppy:  Changes c-calls API to accept client-callback for
3735    allocating extra stack space.
3736    me: Corresponding changes to mlriscGen (using a dummy argument that
3737        does not change the current behavior).
3738    
3739    ----------------------------------------------------------------------
3740    Name: Matthias Blume
3741    Date: 2002/01/28 12:00:00
3742    Tag: Release_110_38
3743    Description:
3744    
3745    This time for real!!!
3746    
3747    ----------------------------------------------------------------------
3748    Name: Matthias Blume
3749    Date: 2002/01/28 10:56:00 EST
3750    Tag: blume-20020128-retraction
3751    Description:
3752    
3753    0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
3754       with blume-Release_110_38-retracted.)
3755    
3756    1. Fixed a problem with incorrect rounding modes in real64.sml.
3757       (Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
3758    
3759    2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
3760       and enums fixed.  The naming of corresponding ML identifiers should
3761       now be consistent again.
3762    
3763    ----------------------------------------------------------------------
3764    Name: Allen Leung
3765    Date: 2002/01/27
3766    Tag: leunga-20020127-nowhere
3767    Description:
3768    
3769       Added a target called nowhere in the configuration scripts.
3770       Enabling this will build the MLRISC 'nowhere' tool (for translating
3771       programs with where-clauses into legal SML code) during installation.
3772    
3773    ----------------------------------------------------------------------
3774    Name: Matthias Blume
3775    Date: 2002/01/25 21:27:00 EST
3776    Tag: blume-Release_110_38-retracted
3777    Description:
3778    
3779    Call it a (working) release!  Version is 110.38. Bootfiles are ready.
3780    
3781    README will be added later.
3782    
3783    !!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
3784    (Release_110_38) removed.  Reason: Last-minute bug fixes.
3785    
3786    ----------------------------------------------------------------------
3787    Name: Matthias Blume
3788    Date: 2002/01/25
3789    Tag: blume-20020125-ffi
3790    Description:
3791    
3792    A large number of tweaks and improvements to ml-nlffi-lib and
3793    ml-nlffigen:
3794    
3795       - ML represenation types have been streamlined
3796       - getter and setter functions work with concrete values, not abstract
3797         ones where possible
3798       - ml-nlffigen command line more flexible (see README file there)
3799       - some bugs have been fixed (hopefully)
3800    
3801    ----------------------------------------------------------------------
3802    Name: Lal George
3803    Date: 2002/01/24
3804    Tag: george-20020124-risc-ra-interface
3805    Description:
3806    
3807       There is a dramatic simplification in the interface to the
3808       register allocator for RISC architectures as a result of making
3809       parallel copy instructions explicit.
3810    
3811    ----------------------------------------------------------------------
3812    Name: Matthias Blume
3813    Date: 2002/01/22
3814    Tag: blume-20020122-x86-ccalls
3815    Description:
3816    
3817    Bug fix for c-calls on x86 (having to do with how char- and
3818    short-arguments are being handled).
3819    
3820    ----------------------------------------------------------------------
3821    Name: Matthias Blume
3822    Date: 2002/01/21
3823    Tag: blume-20020121-ff
3824    Description:
3825    
3826    Another day of fiddling with the FFI...
3827    
3828    1. Bug fix/workaround:  CKIT does not complain about negative array
3829       dimensions, so ml-nlffigen has to guard itself against this possibility.
3830       (Otherwise a negative dimension would send it into an infinite loop.)
3831    
3832    2. Some of the abstract types (light objects, light pointers, most "base"
3833       types) in structure C are now eqtypes.
3834    
3835    3. Added constructors and test functions for NULL function pointers.
3836    
3837    ----------------------------------------------------------------------
3838    Name: Matthias Blume
3839    Date: 2002/01/18
3840    Tag: blume-20020118-ready-for-new-release
3841    Description:
3842    
3843    Made config/srcarchiveurl point to a new place.  (Will provide boot
3844    files shortly.)
3845    
3846    Maybe we christen this to be 110.38?
3847    
3848    ----------------------------------------------------------------------
3849    Name: Matthias Blume
3850    Date: 2002/01/18
3851    Tag: blume-20020118-more-ffifiddle
3852    Description:
3853    
3854    Today's FFI fiddling:
3855