Home My Page Projects Code Snippets Project Openings SML/NJ
Summary Activity Forums Tracker Lists Tasks Docs Surveys News SCM Files

SCM Repository

[smlnj] Diff of /sml/trunk/NOTES/HISTORY
ViewVC logotype

Diff of /sml/trunk/NOTES/HISTORY

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

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