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